Pre-service Physical Sciences Teachers’ Views on Integrating ChatGPT into Teaching: A Case Study

Pre-service Physical Sciences Teachers’ Views on Integrating ChatGPT into Teaching: A Case Study | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Research Article Pre-service Physical Sciences Teachers’ Views on Integrating ChatGPT into Teaching: A Case Study Samuel Jere, Rebecca Bessong, Mamotena Mpeta, Ndanganeni Florence Litshani This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-4783356/v1 This work is licensed under a CC BY 4.0 License Status: Posted Version 1 posted You are reading this latest preprint version Abstract The emergence of artificial intelligence, exemplified by generative chatbots like ChatGPT, has elicited optimism among some educators regarding enhanced teaching and learning methods. Simultaneously, it has raised concerns among others who perceive these chatbots as disruptive to established pedagogical norms developed over centuries. This study investigated and analysed pre-service teachers' perceptions regarding integrating ChatGPT into teaching physical sciences at a rural university. A case study research design that used a qualitative approach was used to collect, analyse and interpret data. This methodology was employed to gain a comprehensive insight into the viewpoints held by physical science pre-service teachers. The study explored the benefits and potential challenges of incorporating emerging technologies like ChatGPT into teaching physical sciences. The theoretical framework that guided the study was the technological, pedagogical content knowledge. Eleven purposively sampled pre-service physical science teachers participated in semi-structured interviews. The collected data were analysed using thematic analysis. The research findings were that ChatGPT has the potential to contribute to teaching physical sciences in lesson planning, preparation, presentation and formative assessment. However, the study revealed that the inability of ChatGPT to answer some questions in physical sciences was of great concern. These findings shed light on how artificial intelligence generative chatbots can be incorporated into science teaching and learning. The findings provide insights for policymakers, science educators and researchers to deepen their understanding of the role of emerging technologies in science education. Artificial Intelligence and Machine Learning Artificial intelligence ChatGPT physical sciences technological pedagogical content knowledge Figures Figure 1 INTRODUCTION ChatGPT was developed by OpenAI and released to the public in November 2022 (Baidoo-Anu & Ansah, 2023; Grassini, 2023). It is an artificial language model based on the transformer architecture, which generates human-like text responses to various prompts and questions (Sharma & Yadav, 2022). It has the capacity to engage in conversational interactions with users. The acronym ChatGPT stands for Chat Generative Pre-trained Transformer. This artificial intelligence chatbot was developed by researchers, engineers and experts in artificial intelligence. Some of the founders of OpenAI and ChatGPT include Elon Musk, Sam Altman, Ilya Sutskever and Greg Brockman (Sharma & Yadav, 2022). The emergence and widespread dissemination of AI tools like ChatGPT can only be compared to other groundbreaking technologies that reshaped human existence, such as gunpowder, Guttenberg’s press and electricity (Wark & Ally, 2020). The impact of AI tools on education and many other areas of human life is yet to be fully understood (Grassini, 2023). They can potentially transform education in ways that not many could have imagined in the past few decades (Cooper, 2023). This implies that new paradigms, theories and strategies are required to cope with the changing educational landscape (Lee, 2023). AI chatbot tools such as ChatGPT mimic the intellectual abilities of human beings and can be used in science education as teaching assistants (Gill et al., 2024; Nguyen et al., 2022; Wollny et al., 2021). Thus, they can be used by teachers to create learning conditions that offer learners guidance, assistance and feedback (Chiu et al., 2023). They facilitate the creation of learning environments that stimulate critical thinking by providing engaging and interactive activities to learners (Grassini, 2023; Nguyen et al., 2022). Chatbots are learning aids used in education to support the teaching of content and skills, improve efficiency in teaching and learning and increase the learners’ motivation to learn (Wollny et al., 2021). Extant literature is awash with the shortcomings of traditional instruction, such as the assumption that all learners are the same and the teacher offering a curriculum that is the same for all at the same time (Dervić et al., 2018; Yavuz, 2020). AI chatbots can create more differentiated, personalised learning conditions than ever (Chiu et al., 2023; Grassini, 2023; Lee, 2023). When integrating AI chatbots in instruction, the teacher encourages learners to generate prompts based on what they do not understand to obtain clarity individually. Moreover, learners are guided to use prompts to create individualised formative assessments, and the chatbots can use previous prompts to provide questions suitable for an individual's level of understanding (Grassini, 2023). The teacher can now honestly be a facilitator of learning that educational theories of learning, such as the constructivist theory, have promoted, thus bridging the gap between theory and practice (Barak, 2017). The teacher can use the AI chatbots to seek ideas to create engaging and motivating lessons. These chatbots can be used as a starting point to find teaching resources (Deng et al., 2023). The teacher can use the AI chatbot to generate resources such as simulations, games, and videos suitable for teaching a particular topic at a given grade level in science education. The teacher then visits the suggested websites and, as a professional, evaluates these resources and decides to include them in instruction. The use of chatbots also aids the teacher in content development and lesson planning (Grassini, 2023; Nguyen et al., 2022). The chatbots offer ideas for producing informative and engaging lesson plans. The teacher can adapt and modify the lesson plan generated by AI in a few moments, thus significantly reducing the teacher's workload (Lee, 2023). Rationale for the Study and Research Questions Despite AI's many potential benefits, these tools have not yet been widely adopted in science education. How teachers pedagogically use AI technologies and their roles in classroom learning remain unclear (Chiu et al., 2023; Lee, 2023). Moreover, teachers resist adopting these teaching tools, while others are unaware of their potential (Nguyen et al., 2022). Furthermore, little is known about the integration of ChatGPT in teaching school learners and in initial teacher preparation (Baidoo-Anu & Ansah, 2023). Therefore, this study sought to determine the extent to which physical sciences preservice teachers understand AI, as exemplified by ChatGPT’s potential benefits and limitations in teaching physical sciences. The study intended to answer the following research questions: 1. To what extent are physical sciences preservice teachers prepared to integrate ChatGPT to develop effective, innovative, adaptive teaching methods? 2. What are potential areas in initial teacher preparation that need improvement to adequately prepare future generations of physical sciences teachers to integrate AI tools like ChatGPT in teaching? The rapid advancement and development of AI tools in the past few years offer opportunities and challenges to their integration into science teaching (West et al., 2023). Understanding how future science teachers are prepared to integrate these tools into their teaching practices is crucial for developing suitable initial teacher preparation courses. This study attempts to close the gap in the literature on how AI tools like ChatGPT can be integrated into the teaching of physical sciences. By investigating the readiness of preservice teachers towards using AI in education, the study contributes to a nuanced understanding of the effectiveness of integrating these emerging technologies into teaching practices and preparing future teachers for a rapidly changing technological landscape. LITERATURE REVIEW Integrating AI in Physical Sciences Teaching While initially, there were attempts to discourage the use of AI chatbots in educational institutions, the debate has since shifted from whether to use AI tools to how they should be used (Deng et al., 2023). Efforts to ban chatbots, such as ChatGPT, in some educational institutions in countries such as the USA (Elsen-Rooney, 2023) are fruitless, as at the time of writing this article, AI tools are already being incorporated into many websites and social media, such as Meta AI in WhatsApp and traditional search engines such as Copilot in Microsoft Edge and Gemini in Chrome. Banning chatbots such as ChatGPT is akin to banning the internet, which is unviable. Moreover, it is now even suggested that the use of AI tools such as ChatGPT will be a necessary professional skill in the future (West et al., 2023). Given these shifting trends in science education, it is essential to explore how science teachers should consider integrating AI tools, such as ChatGPT, into their practices. Researchers acknowledge that AI tools such as ChatGPT can be used to improve teachers’ pedagogical approaches and the learning experiences they offer learners (Grassini, 2023; Lee, 2023). These tools can be used to provide personalised tutoring, automated essay grading, language translation and interactive and adaptive learning (Baidoo-Anu & Ansah, 2023). In assessment, AI tools can be used to automate the grading of learners, which can assist in speedily identifying learning gaps and instituting intervention measures to address them (Grassini, 2023). If grading is automated, it frees the teacher’s time, allowing more focus to be placed on creating innovative lesson plans and providing individualised learning (Grassini, 2023). The ability of AI tools such as ChatGPT to translate educational information from one language to another is considered to have the potential to provide inclusive learning opportunities to many learners whose home language differs from the language of instruction (Baidoo-Anu & Ansah, 2023). Learners with limited proficiency in the language of instruction can use AI tools to overcome language barriers, enhancing their conceptual understanding of the academic content. This can be used to enhance the teaching of physical sciences in economically disadvantaged communities, where the language of instruction is often not the same as the learners’ home language. Teachers are encouraged to integrate AI tools such as ChatGPT in their classroom practices to create innovative pedagogical approaches and seek appropriate content materials and interactive learning activities (Grassini, 2023). Grassini (2023) p. 9 came to the conclusion that “incorporating these AI tools within teacher training programs can equip the next generation of educators with the knowledge and skills to utilise these technologies optimally in their classrooms” . These future teachers should develop the skills to create prompts aligned with their curricula that they ask their learners to use to seek clarity from ChatGPT in the complex concepts they may be learning (Baidoo-Anu & Ansah, 2023). It is acknowledged that science concepts are abstract, complex and difficult for learners to understand. Integrating ChatGPT during teaching implies that the teacher can provide an exploratory phase during instruction. In this exploratory phase, learners can ask ChatGPT questions through appropriate prompts about what they do not understand to provide extra support to what their teacher would have taught them (Baidoo-Anu & Ansah, 2023). Using ChatGPT in initial teacher training can potentially enhance student teachers’ creativity. For example, Liu et al. (2023) found that student teachers using ChatGPT exhibited a higher level of creative ability and better performance than those who did not. Creativity is a critical skill for preservice teachers as they should use it in designing innovative and inspiring lesson plans, finding teaching and learning materials that enhance learning by appealing to the various learning styles of the learners and designing suitable assessment tasks. In all these cases, the integration of ChatGPT can play a pivotal role. However, there are inherent limitations in AI technologies such as ChatGPT, which preservice and practising teachers should be aware of. One of the critical limitations that has been observed is that, in some cases, ChatGPT can provide incorrect or irrelevant information (Lin, 2023). Therefore, when integrating ChatGPT in teaching physical sciences, learners must be made aware of this. ChatGTP should be used to supplement traditional sources of information such as textbooks or other learning resources rather than to replace these sources of information (Lin, 2023). Teachers should guide learners in creating specific prompts to avoid receiving irrelevant or incorrect responses from ChatGPT. If learners get confused by the response, it is often necessary to rephrase their query. ChatGPT will clarify its response. Therefore, when integrating ChatGPT in learning physical sciences, the teacher should teach learners how to create prompts that enhance obtaining relevant information. Other issues regarding using generative AI tools in education are currently unresolved. These issues concern plagiarism and the ethical use of these tools in academic work. Theoretical Framework The study was guided by Mishra and Koehler’s (2006) technological, pedagogical content knowledge (TPCK) framework. Mishra and Koehler (2006) extended Shulman's (1987) work on pedagogical content knowledge (PCK) to include technological knowledge after observing the rapid acceleration of digital technology and its effects on many aspects of human endeavours, including education. Shulman (1987) defined content knowledge as knowledge of facts, concepts, and structure of the subject that a teacher is teaching, including its fundamental principles. At the same time, pedagogy is the broad principles and strategies of classroom management and organisation. During the process of preparing educators, it was observed that subject matter knowledge and teaching strategies were considered distinct components. This observation led Shulman in 1987 to introduce the concept of pedagogical content knowledge. This concept is recognised as “the blending of content and pedagogy into an understanding of how particular topics, problems or issues are organised, represented, and adapted to the diverse interests and abilities of learners and presented for instruction” Shulman (1987). This means that it is important for teachers to possess knowledge of what to teach and how to teach learners. After realising that the rapid changes in the technological landscape were impacting the learning-teaching process, Mishra and Koehler (2006) proposed that technological knowledge was critical to teachers’ knowledge, and thus, the TPCK was formulated. TPCK is described as “… the basis of good teaching with technology and requires an understanding of the representation of concepts using technologies; pedagogical techniques that use technologies in constructive ways to teach content; knowledge of what makes concepts difficult or easy to learn and how technology can help redress some of the problems that students face” Mishra and Koehler (2006). From the TPCK framework, physical sciences teachers require not just knowledge of what to teach and how to teach but also how the content they teach can be enhanced by using digital technology and the teaching strategies by which this can be achieved. METHODOLOGY Research Context The participants were part of a final year Bachelor of Education Honours degree class that took a module on the Methodology of Physical Sciences. In that module, one of the topics was integrating digital technology in teaching sciences, and ChatGPT was used as one of the emerging technologies. The instructor discussed the benefits and limitations of ChatGPT in educational contexts. The participants were instructed on designing prompts they could use in brainstorming in various situations, such as designing lesson plans and searching for appropriate learning/teaching resources such as multimedia, simulations, and virtual experiments. In this context, examples of prompts they would use were: o Design a lesson plan utilising the 5E instructional approach to teach grade 10 learners the topic of electric circuits. o Suggest a formative assessment task that can be used in teaching this topic. o Suggest suitable teaching/learning resources such as simulations or virtual experiments. While the instructor relied on traditional sources of information such as research articles and textbooks to expose the student teachers to the theories underpinning the methodologies of teaching physical sciences, the student teachers were encouraged to use ChatGPT to supplement their understanding. When they found useful information, they could adapt it to the curriculum they were being trained to teach. It was made clear that it is unethical to copy the responses from ChatGPT and present them as their own work. They had to adapt these responses to their own situations. This way, ChatGPT was used to supplement, enrich and complement the module's teaching. Sampling Participants were recruited using the purposive sampling technique to select participants who had shown greater interest and understanding of the integration of ChatGPT in science teaching. The participants were informed of the purpose of the study, and the researchers sought their informed consent, which they granted. The participants were also informed that they were free to withdraw from the study at any time without any adverse effects to them. They were also informed that their responses would be used for academic purposes only and that their identities would not be revealed to anyone. Their responses would remain anonymous. Eleven participants agreed to take part in the study. The data of the participants are summarised in Table 1. The names appearing in Table 1 are pseudonyms. Table 1 Data of the Participants Name of Participant Gender Age John M 26 Bennie M 25 Mary F 27 Elizabeth F 25 Irene F 25 Eva F 23 Asante M 26 Rosemary F 30 Peter M 25 Rodrick M 24 Lebo F 42 Data Collection Semi-structured interviews were used in data collection. In conducting semi-structured interviews in qualitative research, the researcher uses open-ended questions to construct an interview guide (Magaldi & Berler, 2020). During the interview, the researcher uses probing follow-up questions to obtain an in-depth understanding of the phenomena from the participant’s viewpoint. This was the approach used in this study. The following interview guide was used to conduct the interviews: Part A: General Questions o What do you understand by AI, like ChatGPT? o In your own words, what is ChatGPT? Part B: Use of ChartGPT in Lesson Preparation and in Learning and Teaching o How can ChatGPT be used in teaching physical sciences? o How would you use ChatGPT in the following: · Lesson planning, preparation and presentation · Assessment? · Ensuring learners produce prompts suitable to get useful responses from ChatGPT? Part C: Limitations of ChatGPT o ChatGPT can produce inaccurate information. How can you, as a teacher, deal with this? o What are some of the limitations of ChatGPT? Part D: Preservice Teachers’ Perceptions of Integrating ChatGPT in Teaching Physical Sciences o Based on your understanding of the use of ChatGPT in teaching Physical sciences, how comfortable or prepared are you to integrate it into your teaching? o What areas in your learning of using ChatGPT in teaching require improvement? o What can be done to improve preservice teachers’ use of ChatGPT in teaching physical sciences? Semi-structured interviews were conducted using the above interview guide. An electronic device was used to record the participants’ responses to the interview questions after obtaining their permission. Each interview lasted between 30 to 40 minutes. After the interviews, the participants’ responses were transcribed verbatim using software. The participants were allowed to read their transcribed interviews. The participants agreed that the transcribed interviews represented their views. After authentification by member checking, the interview transcripts were ready for analysis. DATA ANALYSIS Data analysis was done using computer-assisted qualitative data analysis software (CAQDAS), ATLAS.ti 8. We relied on Braun and Clarke (2006) six recursive stages of qualitative thematic analysis viz - getting familiar with the data; codes generation; identifying themes; reviewal of themes; defining and naming themes and report writing. The purpose of thematic analysis was to identify, analyse, report and interpret patterns within the data (Braun & Clarke, 2006). The transcribed interview transcripts were imported to ATLAS.ti 8, and the software was used to search for codes within the interview transcripts. Similar codes were then placed in groups to develop sub-themes and themes. RESULTS AND DISCUSSION The themes that emerged from the data are used to organise the results of this study in this section. These results are discussed in the context of the literature reviewed. Preservice Teachers’ Understanding of Artificial Intelligence and ChatGPT Two sub-themes, respectively, represented the preservice teachers’ understanding of AI and ChatGPT: o AI as machines and computers mimicking human intelligence o CahtGPT: AI based on GPT - a language model. The preservice teachers understood AI technology as machines or computer programmes simulating human intelligence. This is what Bennie had to say when asked what he understood by AI: … the theory and development of computer systems capable of performing human tasks or tasks requiring human intelligence, Bennie. Similar sentiments were echoed by Elizabeth and Irene, who described AI as follows: …. involves technology that tries to simulate human intelligence by using machines, Elizabeth. …. the use of machines or computer systems to do what humans can or solve problems that can only be solved by a human, Irene. The preservice teachers’ understanding of ChatGPT as AI based on GPT - a language model is demonstrated by the following quotes: Okay, ChatGPT is a chatbot … generative pre-trained transformer. Looking at what it stands for, I'll say people train it; it's trained by humans to converse with the user in a human-like form, Eva. I understand that it stands for Generative Pre-trained Transformer, and you give it prompts, and then it responds to you like a human being, but it's a computer program, Lebo. The above quotes suggest that the preservice teachers understood artificial intelligence, ChatGPT, and their capabilities. Reiss (2021) claim that definitions of AI have evolved over time as new discoveries were made and see AI as an inanimate matter exhibiting some form of intelligence similar to human intelligence, as suggested by the preservice teachers. This includes machines such as robots performing tasks such as medical diagnosis, and software programs called bots mimicking human intelligence (Reiss, 2021). The understanding of the preservice teachers of ChatGPT as a pre-trained transformer, or a natural language model, trained on a huge amount of words (billions) to be able to respond to text inputs is closely aligned with literature (Sharma & Yadav, 2022). Lesson Preparation, Planning and Presentation The preservice physical science teachers believed that ChatGPT could be helpful in lesson preparation, planning and presentation. Under this theme, their responses were placed in the following sub-themes: lesson objectives, official curriculum documents, lesson preparation, planning, presentation, summarising content, ChatGPT as a resource for teaching/learning materials, simulations, ChatGPT as a tool for learner engagement and ChatGPT as a tool for assessment. Figure 1 demonstrates the relationships between these sub-themes. Regarding lesson objectives, participant Eva had this to say: When ChatGPT gives you the objectives, you, as a teacher, must check them and critically evaluate them. Then, after you evaluate them, you go to a CAPS document or ATP, and then you combine them so that you can provide or come up with the proper lesson plan, Eva. Participants were of the view that curriculum documents must be consulted in lesson preparation, planning, and presentation. As mentioned by Eva, the objectives in any topic that the teacher is teaching while integrating ChatGPT must be verified against curriculum documents. In the case of this group of students, the curriculum documents were the Curriculum and Assessment Policy Statement (CAPS), the Annual Teaching Plans (ATPs) and examination guidelines. Regarding official curriculum documents, Asante had this to say: As a teacher, I will always use other sources to verify my information, to check if the kind of information I am getting from ChatGPT is appropriate or the kind of information I want. Not just copying everything from ChatGPT. I will use the CAPS document and the different textbooks to check if it’s the correct information or if it's wrong, Asante. In terms of lesson preparation , this is what Lebo had to say: For lesson preparation, I think mostly to assist me with teaching aids because then it will direct me as to where I can find the simulations for the hands-on experiments and also give me ideas on what type of experiments I can perform with learners for them to understand that particular concept that I want to teach about , Lebo. Other participants, just like Lebo, mentioned that ChatGPT was useful in searching for resources such as simulations and videos, websites and summarising content. Under the sub-theme of assessment in physical sciences , Mary said: It can help teachers who teach physical sciences know how to use ChatGPT to assess learners or create questions they can use in their lessons. In terms of assessment, I can use it to generate questions, maybe like multiple choice, true or false questions, long and short questions, Mary. Under the sub-theme of ChatGPT as a tool for learner engagement , Irene said: ChatGPT can help engage the learners. The students learn better when they're interacting in class. If I can give a question to learners and ask them to please check this information from ChatGPT, it will help learners be active and engaged in the classroom , Irene. Other participants expressed similar sentiments, indicating that preservice teachers regarded ChatGPT as a useful tool for enhancing learner engagement and generating assessment tasks and teaching/learning resources. The sentiments expressed by the preservice teachers on the potential benefits of integrating ChatGPT in teaching physical sciences align with findings from previous studies. For example, a study by Jo (2024) revealed that using ChatGPT in learning results in enhanced knowledge acquisition and application, with the novelty associated with personalised learning using ChatGPT stimulating engagement, thus making learning exciting and motivating. The preservice teachers' perceptions that ChatGPT is a useful tool for generating questions or responding to learners’ queries align with the findings from the literature. A study by Lee and Zhai (2024) revealed that preservice science teachers regarded ChatGPT as useful in learners' formative assessment. This is supported by Rahman and Watanobe (2023) who claim that physics teachers can use ChatGPT to generate practice questions, quizzes and explanations for formative assessment. They also argue that ChatGPT can provide step-by-step solutions, thereby developing learners’ problem-solving skills. Rahman and Watanobe (2023) also reported that ChatGPT is a useful tool in lesson planning in chemistry and physics, as indicated by the preservice teachers. This statement corresponds with the research findings indicating that ChatGPT is capable of creating effective lesson plans based on the 5E model (Cooper, 2023). Prompt Engineering The participants' responses on prompt engineering reflect that they had an appreciable level of understanding. When asked to describe how they could assist their learners to ask prompts in ChatGPT, some of their responses were as follows: Because, you see, your students are in grade 10, grade 11 or 12, and the prompts they give must be specific for that level of education. Because ChatGPT can sometimes provide information that is too advanced for the students. So, on the ChatGPT, you need to ask a specific question related to that grade or that standard so that it can provide you with the correct information. And you need to ask for specific information so that it won't give you the wrong information or the more advanced information to that standard, Rosemary. Let's say we are doing electrical circuits, and they have a question about how resistors in parallel affect the current. They must also be specific that we are doing it for grade 10. The content needs to be for a grade 10 learner so that then when it gives out the answers, the answers will be in such a way that it's understandable to a learner that is in grade 10 rather than maybe an electrician who then has more knowledge about electricity than a grade 10 learner, Lebo. I can ask my learners to ask clear, specific questions. If possible, they must also provide context and follow-up questions, Peter. The responses of the participants echo some of the features of prompt engineering that are discussed in Giray (2023), who defines a prompt as specific instruction or a query that a learner provides to ChatGPT. The four elements of effective prompting identified by Giray (2023) are: instruction, context, input data and output indicator. As noted by preservice teachers, the instruction has to be specific to guide the chatbot to an appropriate response. The context is the additional information that the learner provides in the prompt to guide ChatGPT to generate an accurate output (Giray, 2023). Peter stated that learners have to be taught to provide context, and other participants, such as Rosemary, stated that learners have to be specific and that they need responses suitable for the grade 10 level of education, for example. Regarding input data, Giray (2023) describes this as the core of the prompt and is the specific question that ChatGPT should respond to. This is discussed by Lebo and Rodrick as follows: I think I will have to train or assist my learners in properly putting the prompts so that they give the relevant information. Because if you don't give it the correct prompts, it will just give you any information, seeing that it has access to a lot of information. So, learners need to know that there are certain words that they need to use that will assist them in getting specific answers from ChatGPT , Lebo. We must be able to teach our learners how to prompt ChatGPT so that they can get problems, explanations, and summaries on topics we've taught them , Rodrick. As for the output indicator, the learner must direct ChatGPT regarding the response format required (Giray, 2023). The learner can guide ChatGPT about the length of the response and the details of the response sought. While Rodrick stated the need to specify the type of response, such as problems, summary or explanation, the participant fell short of stating that learners must describe the length of the response. Limitations of ChatGPT and Suggestions on Mitigating the Limitations The major limitation that ChatGPT currently possesses, which was of great concern to preservice physical sciences teachers, was that sometimes ChatGPT provided inaccurate information (Elmas et al., 2024). While the teachers claimed they could notice inaccurate information, they feared that their learners would be unable to differentiate between accurate and inaccurate information. Some of the quotes from the participants are as follows: As ChatGPT can provide inaccurate information, as a teacher, I will always use other sources to verify my information and check if the kind of information I am getting from ChatGPT is appropriate or the kind of information I want. Not just copying everything from ChatGPT. I will use the CAPS document and the different textbooks to check if that's the correct information or if it's wrong , Asante. When I'm teaching my learners, when I get the information from ChatGPT, I have to go through it, and I have to consult some sources before I go and teach my learners such information, and I have to use it as an aid , Bennie. The teacher must emphasise the limitations of using ChatGPT, as many people or learners don't know where they have to stop. They take all the information, so they must have the knowledge of the limitations. We need to know a lot about limitations , Elizabeth. I can always teach students that they have to learn a good way of putting in the prompt questions, and they don't have to rely too much on it. They have to take it as an assistant , John. The teachers were of the view that using ChatGPT in conjunction with traditional sources of information would mitigate some of its constraints. However, while they claimed that they would do this, it was not clear if learners would also manage to utilize ChatGPT in this manner as refelected by the following quotes: As a teacher, on my side, I think I have an advantage because I already have the content knowledge. So, I'll be able to identify that this information is incorrect and then use other resources, be it a textbook or other resources, to get the correct information . For me, it starts being dangerous on the part of the learner because the learner is not equipped enough to realise that this information is incorrect. And looking at our children these days, they are over-reliant on digital information. So that's where it will be a challenge as a teacher to encourage my learners to get proper content; they need not rely on ChatGPT only. They also need to refer to the proper textbook we use for physical sciences to get the correct information, Lebo. A teacher should understand the content so that we can see the flaws in the answers that are out there. Sometimes, it's not just ChatGPT; even the learners' guides sometimes have the wrong information. So, a teacher should be able to understand the content. ChatGPT, though it's worse than the guides that the learners sometimes use, requires a teacher to understand the subject matter and the subject content. A teacher must understand the curriculum document. A teacher must understand the exam guidelines so that we can see the flaws in the answers from ChatGPT or explanations , Rodrick. The constraints identified by the preservice educators, such as the presence of inaccurate responses, are likely a result of the existing capabilities of chatbots. A study by Fergus et al. (2023) revealed that in chemistry, ChatGPT can satisfactorily provide answers to questions at the remembering and comprehension levels in Bloom’s taxonomy and struggles with questions that require application or interpretation. They particularly noted that the chatbot faces difficulties with nontext information, such as questions with graphical or numeric information. Additionally, comparable findings showed that ChatGPT was unable to achieve over 37% accuracy in responses to multiple-choice questions or brief answers within an introductory-level university chemistry course (Leon & Vidhani, 2023). This means that learners would get incorrect responses most of the time when they attempted to use ChatGPT. Poor performance in responding to biology questions was also observed by Elmas et al. (2024). Another potential limitation of ChatGPT identified by preservice teachers was that learners could develop over-reliance on it. There was fear that ChatGPT could assume the role of an epistemic authority for learners despite the limitation of providing inaccurate information in some chemistry and physics problems. Some of the quotes illustrating this are: But it's (ChatGPT) also not good at all for our learners. As we know, learners are much lazier when searching for information. They will rely much more on ChatGPT as we show them how to use it, Rosemary. The people (learners) using it need to use it responsibly. Science is not just about getting a correct answer but how you get to that answer. So, in terms of developing critical skills and problem-solving, that still needs to be done by a human being. So, people need not over-rely on it to provide the correct answer and not understand how to get to that correct answer , Lebo. Preservice teachers believed that while ChatGPT could be leveraged to provide science information, traditional sources of information should be used to verify that information. They noted that this would be difficult for learners. Yes, and I'll also advise them that when they get information on ChatGPT, they must not take it as it is, but they must look for another source of information and combine it with that information from ChatGPT , Rosemary. For me, it starts being dangerous on the part of the learner because the learner is not equipped enough to realise that this information is incorrect. And looking at our children these days, they are over-reliant on digital information. So that's where it will be a challenge as a teacher to encourage my learners to get proper content; they need not rely on ChatGPT only. They also need to refer to the proper textbook we use for physical sciences to get the correct information , Lebo. These results are aligned with the literature which suggests that developers of ChatGPT should make an effort to ensure that their algorithms produce more accurate responses in education to prevent harm to learners (Cambra-Fierro et al., 2024). While empirical studies indicate that ChatGPT has many potential benefits at different levels of education ranging from increasing the teachers’ well-being and reducing stress and assisting in the achievement of learning objectives and creation of effective lesson plans (Cambra-Fierro et al., 2024; Wijaya et al., 2024), there is a serious need for teachers to be aware of the limitations discussed by the preservice teachers. Areas of Improvement The preservice teachers identified auto-marking learners’ works, preventing plagiarism and how teachers can be assisted in dealing with inaccurate responses as areas that needed improvement in preparing preservice teachers to integrate ChatGPT in teaching Physical Sciences. The following quotes exemplify their sentiments: We are not quite sure about the use of ChatGPT to assess learners' work. I think they are not being taught enough. So the teachers do not emphasise them. Limitations. I think it's the same as the limitations. But it has got limitations. Such as providing incorrect responses. And I’m unsure how to deal with that in your teaching practice , Mary. The area that must be improved is how learners can avoid plagiarism from ChatGPT. In higher education, most assessments are done online, and learners use ChatGPT as a source of plagiarism. I think they need to come up with solutions I'm not sure if there can be apps that can be used to help to avoid plagiarism , Peter. The area that I did not find well was that when writing assignments, ChatGPT makes it too easy for a person to write an assignment; you may end up not learning that much because you rely on ChatGPT; you only need a summary, and then you can go to ChatGPT to have an idea after having an idea of what you are searching on ChatGPT can assist you. Therefore, it will limit the critical thinking of students , Irene. The participants' responses imply that preservice teachers must develop their skills in automatic grading of student work and explore how well they can incorporate these techniques into their teaching practices. The physical sciences preservice teachers indicated that they were very keen to incorporate ChatGPT in their teaching practices, as expressed by Rodrick in the following quote: Just incorporating more on the use of AI. We are in the 4IR at this time, and technology is taking over. So, we cannot be left behind as physical science teachers. Upcoming physical science teachers must be taught how to use AI in their teaching. How can they use technology in their teaching? Rodrick. The preservice teachers’ perceptions regarding the benefits of integrating ChatGPT in learning physical sciences, such as explaining complex science concepts, assisting in problem-solving and conceptual understanding, and increasing the learners’ motivation, interest and engagement, are supported by literature (Iyamuremye & Ndihokubwayo, 2024; Kodkin & Artem’eva, 2024; Taani & Alabidi, 2024). However, it is important that teachers should take cognisance of the limitations of ChatGPT, particularly in responding to questions that require application and solving complex problems in physics and chemistry (Leon & Vidhani, 2023). Conclusion The study has provided empirical evidence for the potential areas in which ChatGPT can be used to enhance physical sciences instruction. These areas include lesson preparation, planning and presentation. In lesson preparation and planning, the preservice teachers believed that ChatGPT is a tool that can be used for content generation, searching for teaching/learning materials, and generating formative assessment tasks. ChatGPT was also found to have the potential to assist in generating innovative lesson plans. Nevertheless, the investigation found that ChatGPT should function in a supplementary capacity. It is advisable to use it to complement conventional sources of information, such as textbooks and formal curriculum guides. It is advisable for the creators of ChatGPT to enhance the accuracy of the generated responses to questions about the physical sciences. Currently, the application often provides incorrect answers, which may contribute to misconceptions in fields such as physics and chemistry. This could adversely affect the education in these areas. Policymakers should ensure that they provide professional development to practising teachers on the potential benefits and constraints of ChatGPT in physical sciences instruction. Further investigation is advised to assess ChatGPT's influence on physical sciences education. Implementing experimental and quasi-experimental research in educational settings could mitigate some of the limitations identified in this research, which was constrained by its case study approach and the small sample size from a rural area. Declarations Acknowledgements The authors would like to express their gratitude to the participants for taking the time to respond to the interviews and to the university research ethics committee for granting them permission to conduct the study. Declaration of Interest The authors declare no conflict of interest. References Baidoo-Anu, D., & Ansah, L. O. (2023). Education in the era of generative artificial intelligence (AI): Understanding the potential benefits of ChatGPT in promoting teaching and learning. Journal of AI , 7 (1), 52-62. https://doi.org/10.61969/jai.1337500 Barak, M. (2017). Science teacher education in the twenty-first century: A pedagogical framework for technology-integrated social constructivism. Research in Science Education , 47 , 283-303. https://doi.org/10.1007/s11165-015-9501-y Braun, V., & Clarke, V. (2006). Using thematic analysis in psychology. Qualitative research in psychology , 3 (2), 77-101. https://doi.org/10.1191/1478088706qp063oa Cambra-Fierro, J. J., Blasco, M. F., López-Pérez, M.-E. E., & Trifu, A. (2024). ChatGPT adoption and its influence on faculty well-being: An empirical research in higher education. Education and Information Technologies . https://doi.org/10.1007/s10639-024-12871-0 Chiu, T. K., Moorhouse, B. L., Chai, C. S., & Ismailov, M. (2023). Teacher support and student motivation to learn with Artificial Intelligence (AI) based chatbot. Interactive Learning Environments , 1-17. https://doi.org/10.1080/10494820.2023.2172044 Cooper, G. (2023). Examining science education in ChatGPT: An exploratory study of generative artificial intelligence. Journal of Science Education and Technology , 32 (3), 444-452. https://doi.org/10.1007/s10956-023-10039-y Deng, J. M., Lalani, Z., McDermaid, L. A., & Szozda, A. R. (2023). Using generative artificial intelligence in chemistry education research: prioritizing ethical use and accessibility. Chamical Education . https://doi.org/10.26434/chemrxiv-2023-24zfl Dervić, D., Glamočić, D. S., Gazibegović-Busuladžić, A., & Mešić, V. (2018). Teaching physics with simulations: Teacher-centered versus student-centered approaches. Journal of Baltic Science Education , 17 (2), 288. https://doi.org/10.33225/jbse/18.17.288 Elmas, R., Adiguzel-Ulutas, M., & Yılmaz, M. (2024). Examining ChatGPT’s validity as a source for scientific inquiry and its misconceptions regarding cell energy metabolism. Education and Information Technologies . https://doi.org/10.1007/s10639-024-12749-1 Elsen-Rooney, M. (2023). NYC education department blocks ChatGPT on school devices, networks . Chalkbeat New York. Retrieved 20 April from https://www.chalkbeat.org/newyork/2023/1/3/23537987/nyc-schools-ban-chatgpt-writing-artificial-intelligence/ Fergus, S., Botha, M., & Ostovar, M. (2023). Evaluating academic answers generated using ChatGPT. Journal of chemical Education , 100 (4), 1672-1675. https://doi.org/10.1021/acs.jchemed.3c00087 Gill, S. S., Xu, M., Patros, P., Wu, H., Kaur, R., Kaur, K., Fuller, S., Singh, M., Arora, P., & Parlikad, A. K. (2024). Transformative effects of ChatGPT on modern education: Emerging Era of AI Chatbots. Internet of Things and Cyber-Physical Systems , 4 , 19-23. https://doi.org/10.1016/j.iotcps.2023.06.002 Giray, L. (2023). Prompt engineering with ChatGPT: a guide for academic writers. Annals of biomedical engineering , 51 (12), 2629-2633. https://doi.org/10.1007/s10439-023-03272-4 Grassini, S. (2023). Shaping the future of education: exploring the potential and consequences of AI and ChatGPT in educational settings. Education Sciences , 13 (7), 692. https://doi.org/10.3390/educsci13070692 Iyamuremye, A., & Ndihokubwayo, K. (2024). Exploring secondary school students’ interest and mastery of atomic structure and chemical bonding through ChatGPT. Educational Journal of Artificial Intelligence and Machine Learning , 1 (1), 1-13. https://doi.org/10.58197/prbl/9hk37296 Jo, H. (2024). From concerns to benefits: a comprehensive study of ChatGPT usage in education. International Journal of Educational Technology in Higher Education , 21 (1), 35. https://doi.org/10.1186/s41239-024-00471-4 Kodkin, V. L., & Artem’eva, E. V. (2024). ChatGPT: Application in Chemistry Education and Challenges. Journal of Computer and Communications , 12 (03), 196-206. https://doi.org/10.4236/jcc.2024.123012 Lee, G. G., & Zhai, X. (2024). Using ChatGPT for Science Learning: A Study on Pre-service Teachers' Lesson Planning. IEEE Transactions on Learning Technologies . https://doi.org/10.1109/tlt.2024.3401457 Lee, H. (2023). The rise of ChatGPT: Exploring its potential in medical education. Anatomical sciences education . https://doi.org/10.1002/ase.2270 Leon, A. J., & Vidhani, D. (2023). ChatGPT needs a chemistry tutor too. Journal of chemical Education , 100 (10), 3859-3865. https://doi.org/10.1021/acs.jchemed.3c00288 Lin, X. (2023). Exploring the role of ChatGPT as a facilitator for motivating self-directed learning among adult learners. Adult Learning . https://doi.org/10.1177/10451595231184928 Liu, Z., Vobolevich, A., & Oparin, A. (2023). The Influence of AI ChatGPT on Improving Teachers' Creative Thinking. International Journal of Learning, Teaching and Educational Research , 22 (12), 124-139. https://doi.org/10.26803/ijlter.22.12.7 Magaldi, D., & Berler, M. (2020). Semi-structured interviews. Encyclopedia of Personality and Individual Differences , 4825-4830. https://doi.org/10.1007/978-3-319-24612-3_857 Mishra, P., & Koehler, M. J. (2006). Technological pedagogical content knowledge: A framework for teacher knowledge. Teachers College Record , 108 (6), 1017-1054. https://doi.org/10.1111/j.1467-9620.2006.00684.x Nguyen, G., Nguyen, N., & Giang, N. (2022). Situation and Proposals for Implementing Artificial Intelligence-based Instructional Technology in Vietnamese Secondary Schools. International Journal of Emerging Technologies in Learning , 17 (18), 53-75. https://doi.org/10.3991/ijet.v17i18.31503 Rahman, M. M., & Watanobe, Y. (2023). ChatGPT for education and research: Opportunities, threats, and strategies. Applied Sciences , 13 (9). https://doi.org/10.3390/app13095783 Reiss, M. J. (2021). The Use of Al in Education: Practicalities and Ethical Considerations. London Review of Education , 19 (1), n1. https://doi.org/10.14324/lre.19.1.05 Sharma, S., & Yadav, R. (2022). Chat GPT–A technological remedy or challenge for education system. Global Journal of Enterprise Information System , 14 (4), 46-51. https://doi.org/10.18311/gjeis/2022 Shulman, L. (1987). Knowledge and teaching: Foundations of the new reform. Harvard Educational Review , 57 (1), 1-23. https://doi.org/10.17763/haer.57.1.j463w79r56455411 Taani, O., & Alabidi, S. (2024). ChatGPT in education: benefits and challenges of ChatGPT for mathematics and science teaching practices. International Journal of Mathematical Education in Science and Technology , 1-30. https://doi.org/10.1080/0020739x.2024.2357341 Wark, N., & Ally, M. (2020). An Emergent Pedagogical Framework for Integrating Emergent Technologies. In Emerging Technologies and Pedagogies in the Curriculum , 89. https://doi.org/10.1007/978-981-15-0618-5_6 West, J. K., Franz, J. L., Hein, S. M., Leverentz-Culp, H. R., Mauser, J. F., Ruff, E. F., & Zemke, J. M. (2023). An analysis of ai-generated laboratory reports across the chemistry curriculum and student perceptions of chatgpt. Journal of chemical Education , 100 (11), 4351-4359. https://doi.org/10.1021/acs.jchemed.3c00581 Wijaya, T. T., Su, M., Cao, Y., Weinhandl, R., & Houghton, T. (2024). Examining Chinese preservice mathematics teachers’ adoption of AI chatbots for learning: Unpacking perspectives through the UTAUT2 model. Education and Information Technologies . https://doi.org/10.1007/s10639-024-12837-2 Wollny, S., Schneider, J., Di Mitri, D., Weidlich, J., Rittberger, M., & Drachsler, H. (2021). Are we there yet?-a systematic literature review on chatbots in education. Frontiers in Artificial Intelligence , 4 . https://doi.org/10.3389/frai.2021.654924 Yavuz, A. C. (2020). The effects of differentiated instruction on Turkish students’ L2 achievement, and student and teacher perceptions. Eurasian Journal of Applied Linguistics , 6 (2), 313-335. https://doi.org/10.32601/ejal.776002 Additional Declarations The authors declare no competing interests. Cite Share Download PDF Status: Posted Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-4783356","acceptedTermsAndConditions":true,"allowDirectSubmit":true,"archivedVersions":[],"articleType":"Research Article","associatedPublications":[],"authors":[{"id":330316670,"identity":"45f18866-1ed5-4a9b-b677-9d1ab3b9e052","order_by":0,"name":"Samuel Jere","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAABAUlEQVRIiWNgGAWjYNCCAgZmIMn4gIFBglgtBmAtzAYkaQEBNuLUy7efTpP4YMDAzi99+Fg17w6LxH7pBsYPPxjq5HGafyZ3m+QMoMMk+9LSbvOekUicOecAs2QPA5thA04n5W6T5gFqMTjDY3abt03C2OBGAoM0AwMPIy4t8v1vt0n/AWqxP8P/rRikxf5GAvNvYDjY49LCcANoCyjEDHh42JiBWuQMJBLYgLYYJOJ02I23my17DCSYJc6wGUvOBWqRuJHYBhRJSMbtsNyNN35U2CTz9zA//PC2rY6Hf0byYaBInS1Oh0GARDISB+RxA/zqQcCOsJJRMApGwSgYsQAA4GpG6sgGfVsAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0002-3112-4036","institution":"University of Venda","correspondingAuthor":true,"prefix":"","firstName":"Samuel","middleName":"","lastName":"Jere","suffix":""},{"id":330317628,"identity":"18343ce7-df05-47f0-bbf0-3d341e45b48e","order_by":1,"name":"Rebecca Bessong","email":"","orcid":"https://orcid.org/0000-0001-9740-4340","institution":"University of Venda","correspondingAuthor":false,"prefix":"","firstName":"Rebecca","middleName":"","lastName":"Bessong","suffix":""},{"id":330318687,"identity":"5fe7ca06-83af-40ba-888d-9766799be883","order_by":2,"name":"Mamotena Mpeta","email":"","orcid":"https://orcid.org/0000-0002-5005-0765","institution":"University of Venda","correspondingAuthor":false,"prefix":"","firstName":"Mamotena","middleName":"","lastName":"Mpeta","suffix":""},{"id":330319309,"identity":"acb50228-4e56-45ca-95bb-0350cd7e7a3a","order_by":3,"name":"Ndanganeni Florence Litshani","email":"","orcid":"https://orcid.org/0000-0002-7370-9657","institution":"University of Venda","correspondingAuthor":false,"prefix":"","firstName":"Ndanganeni","middleName":"Florence","lastName":"Litshani","suffix":""}],"badges":[],"createdAt":"2024-07-22 16:39:44","currentVersionCode":1,"declarations":{"humanSubjects":true,"vertebrateSubjects":false,"conflictsOfInterestStatement":false,"humanSubjectEthicalGuidelines":true,"humanSubjectConsent":true,"humanSubjectClinicalTrial":false,"humanSubjectCaseReport":false,"vertebrateSubjectEthicalGuidelines":false},"doi":"10.21203/rs.3.rs-4783356/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-4783356/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":60911163,"identity":"2088045e-4a09-492b-a496-ec899f19ad52","added_by":"auto","created_at":"2024-07-23 12:53:57","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":142931,"visible":true,"origin":"","legend":"\u003cp\u003eIntegrating ChatGPT into Lesson Preparation, Planning and Presentations\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-4783356/v1/220a235124d9ba60b54523d5.png"},{"id":60912376,"identity":"555ff666-ea1c-4c6c-979c-e6c509ca77e9","added_by":"auto","created_at":"2024-07-23 13:01:57","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":636464,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-4783356/v1/df1ec874-2264-4720-929c-48e989513fb7.pdf"}],"financialInterests":"The authors declare no competing interests.","formattedTitle":"\u003cp\u003e\u003cstrong\u003ePre-service Physical Sciences Teachers’ Views on Integrating ChatGPT into Teaching: A Case Study\u003c/strong\u003e\u003c/p\u003e","fulltext":[{"header":"INTRODUCTION","content":"\u003cp\u003eChatGPT was developed by OpenAI and released to the public in November 2022 (Baidoo-Anu \u0026amp; Ansah, 2023; Grassini, 2023). It is an artificial language model based on the transformer architecture, which generates human-like text responses to various prompts and questions (Sharma \u0026amp; Yadav, 2022). It has the capacity to engage in conversational interactions with users. The acronym ChatGPT stands for Chat Generative Pre-trained Transformer. This artificial intelligence chatbot was developed by researchers, engineers and experts in artificial intelligence. Some of the founders of OpenAI and ChatGPT include Elon Musk, Sam Altman, Ilya Sutskever and Greg Brockman (Sharma \u0026amp; Yadav, 2022).\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe emergence and widespread dissemination of AI tools like ChatGPT can only be compared to other groundbreaking technologies that reshaped human existence, such as gunpowder, Guttenberg’s press and electricity (Wark \u0026amp; Ally, 2020). The impact of AI tools on education and many other areas of human life is yet to be fully understood (Grassini, 2023). They can potentially transform education in ways that not many could have imagined in the past few decades (Cooper, 2023). This implies that new paradigms, theories and strategies are required to cope with the changing educational landscape (Lee, 2023).\u003c/p\u003e\n\u003cp\u003eAI chatbot tools such as ChatGPT mimic the intellectual abilities of human beings and can be used in science education as teaching assistants (Gill et al., 2024; Nguyen et al., 2022; Wollny et al., 2021). Thus, they can be used by teachers to create learning conditions that offer learners guidance, assistance and feedback (Chiu et al., 2023). They facilitate the creation of learning environments that stimulate critical thinking by providing engaging and interactive activities to learners (Grassini, 2023; Nguyen et al., 2022). Chatbots are learning aids used in education to support the teaching of content and skills, improve efficiency in teaching and learning and increase the learners’ motivation to learn (Wollny et al., 2021).\u003c/p\u003e\n\u003cp\u003eExtant literature is awash with the shortcomings of traditional instruction, such as the assumption that all learners are the same and the teacher offering a curriculum that is the same for all at the same time (Dervić et al., 2018; Yavuz, 2020). AI chatbots can create more differentiated, personalised learning conditions than ever (Chiu et al., 2023; Grassini, 2023; Lee, 2023). When integrating AI chatbots in instruction, the teacher encourages learners to generate prompts based on what they do not understand to obtain clarity individually. Moreover, learners are guided to use prompts to create individualised formative assessments, and the chatbots can use previous prompts to provide questions suitable for an individual's level of understanding (Grassini, 2023). The teacher can now honestly be a facilitator of learning that educational theories of learning, such as the constructivist theory, have promoted, thus bridging the gap between theory and practice (Barak, 2017).\u003c/p\u003e\n\u003cp\u003eThe teacher can use the AI chatbots to seek ideas to create engaging and motivating lessons. These chatbots can be used as a starting point to find teaching resources (Deng et al., 2023). The teacher can use the AI chatbot to generate resources such as simulations, games, and videos suitable for teaching a particular topic at a given grade level in science education. The teacher then visits the suggested websites and, as a professional, evaluates these resources and decides to include them in instruction. The use of chatbots also aids the teacher in content development and lesson planning (Grassini, 2023; Nguyen et al., 2022). The chatbots offer ideas for producing informative and engaging lesson plans. The teacher can adapt and modify the lesson plan generated by AI in a few moments, thus significantly reducing the teacher's workload (Lee, 2023).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eRationale for the Study and Research Questions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eDespite AI's many potential benefits, these tools have not yet been widely adopted in science education. How teachers pedagogically use AI technologies and their roles in classroom learning remain unclear (Chiu et al., 2023; Lee, 2023). Moreover, teachers resist adopting these teaching tools, while others are unaware of their potential (Nguyen et al., 2022). Furthermore, little is known about the integration of ChatGPT in teaching school learners and in initial teacher preparation (Baidoo-Anu \u0026amp; Ansah, 2023). Therefore, this study sought to determine the extent to which physical sciences preservice teachers understand AI, as exemplified by ChatGPT’s potential benefits and limitations in teaching physical sciences. The study intended to answer the following research questions:\u003c/p\u003e\n\u003cp\u003e1.\u0026nbsp; \u0026nbsp;To what extent are physical sciences preservice teachers prepared to integrate ChatGPT to develop effective, innovative, adaptive teaching methods?\u003c/p\u003e\n\u003cp\u003e2.\u0026nbsp; \u0026nbsp;What are potential areas in initial teacher preparation that need improvement to adequately prepare future generations of physical sciences teachers to integrate AI tools like ChatGPT in teaching?\u003c/p\u003e\n\u003cp\u003eThe rapid advancement and development of AI tools in the past few years offer opportunities and challenges to their integration into science teaching (West et al., 2023). Understanding how future science teachers are prepared to integrate these tools into their teaching practices is crucial for developing suitable initial teacher preparation courses. This study attempts to close the gap in the literature on how AI tools like ChatGPT can be integrated into the teaching of physical sciences. By investigating the readiness of preservice teachers towards using AI in education, the study contributes to a nuanced understanding of the effectiveness of integrating these emerging technologies into teaching practices and preparing future teachers for a rapidly changing technological landscape.\u003c/p\u003e"},{"header":"LITERATURE REVIEW","content":"\u003cp\u003e\u003cstrong\u003eIntegrating AI in Physical Sciences Teaching\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eWhile initially, there were attempts to discourage the use of AI chatbots in educational institutions, the debate has since shifted from \u003cem\u003ewhether\u003c/em\u003e to use AI tools to \u003cem\u003ehow\u003c/em\u003e they should be used (Deng et al., 2023). Efforts to ban chatbots, such as ChatGPT, in some educational institutions in countries such as the USA (Elsen-Rooney, 2023) are fruitless, as at the time of writing this article, AI tools are already being incorporated into many websites and social media, such as Meta AI in WhatsApp and traditional search engines such as Copilot in Microsoft Edge and Gemini in Chrome. Banning chatbots such as ChatGPT is akin to banning the internet, which is unviable. Moreover, it is now even suggested that the use of AI tools such as ChatGPT will be a necessary professional skill in the future (West et al., 2023). Given these shifting trends in science education, it is essential to explore how science teachers should consider integrating AI tools, such as ChatGPT, into their practices.\u003c/p\u003e\n\u003cp\u003eResearchers acknowledge that AI tools such as ChatGPT can be used to improve teachers’ pedagogical approaches and the learning experiences they offer learners (Grassini, 2023; Lee, 2023). These tools can be used to provide personalised tutoring, automated essay grading, language translation and interactive and adaptive learning (Baidoo-Anu \u0026amp; Ansah, 2023). In assessment, AI tools can be used to automate the grading of learners, which can assist in speedily identifying learning gaps and instituting intervention measures to address them (Grassini, 2023). If grading is automated, it frees the teacher’s time, allowing more focus to be placed on creating innovative lesson plans and providing individualised learning (Grassini, 2023).\u003c/p\u003e\n\u003cp\u003eThe ability of AI tools such as ChatGPT to translate educational information from one language to another is considered to have the potential to provide inclusive learning opportunities to many learners whose home language differs from the language of instruction (Baidoo-Anu \u0026amp; Ansah, 2023). Learners with limited proficiency in the language of instruction can use AI tools to overcome language barriers, enhancing their conceptual understanding of the academic content. This can be used to enhance the teaching of physical sciences in economically disadvantaged communities, where the language of instruction is often not the same as the learners’ home language.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eTeachers are encouraged to integrate AI tools such as ChatGPT in their classroom practices to create innovative pedagogical approaches and seek appropriate content materials and interactive learning activities (Grassini, 2023). Grassini (2023) p. 9 came to the conclusion that \u003cem\u003e“incorporating these AI tools within teacher training programs can equip the next generation of educators with the knowledge and skills to utilise these technologies optimally in their classrooms”\u003c/em\u003e. These future teachers should develop the skills to create prompts aligned with their curricula that they ask their learners to use to seek clarity from ChatGPT in the complex concepts they may be learning (Baidoo-Anu \u0026amp; Ansah, 2023). It is acknowledged that science concepts are abstract, complex and difficult for learners to understand. Integrating ChatGPT during teaching implies that the teacher can provide an exploratory phase during instruction. In this exploratory phase, learners can ask ChatGPT questions through appropriate prompts about what they do not understand to provide extra support to what their teacher would have taught them (Baidoo-Anu \u0026amp; Ansah, 2023).\u003c/p\u003e\n\u003cp\u003eUsing ChatGPT in initial teacher training can potentially enhance student teachers’ creativity. For example, Liu et al. (2023) found that student teachers using ChatGPT exhibited a higher level of creative ability and better performance than those who did not. Creativity is a critical skill for preservice teachers as they should use it in designing innovative and inspiring lesson plans, finding teaching and learning materials that enhance learning by appealing to the various learning styles of the learners and designing suitable assessment tasks. In all these cases, the integration of ChatGPT can play a pivotal role. However, there are inherent limitations in AI technologies such as ChatGPT, which preservice and practising teachers should be aware of.\u003c/p\u003e\n\u003cp\u003eOne of the critical limitations that has been observed is that, in some cases, ChatGPT can provide incorrect or irrelevant information (Lin, 2023). Therefore, when integrating ChatGPT in teaching physical sciences, learners must be made aware of this. ChatGTP should be used to supplement traditional sources of information such as textbooks or other learning resources rather than to replace these sources of information (Lin, 2023). Teachers should guide learners in creating specific prompts to avoid receiving irrelevant or incorrect responses from ChatGPT. If learners get confused by the response, it is often necessary to rephrase their query. ChatGPT will clarify its response. Therefore, when integrating ChatGPT in learning physical sciences, the teacher should teach learners how to create prompts that enhance obtaining relevant information. Other issues regarding using generative AI tools in education are currently unresolved. These issues concern plagiarism and the ethical use of these tools in academic work.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eTheoretical Framework\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study was guided by Mishra and Koehler’s (2006) technological, pedagogical content knowledge (TPCK) framework. Mishra and Koehler (2006) extended Shulman's (1987) work on pedagogical content knowledge (PCK) to include technological knowledge after observing the rapid acceleration of digital technology and its effects on many aspects of human endeavours, including education. Shulman (1987) defined content knowledge as knowledge of facts, concepts, and structure of the subject that a teacher is teaching, including its fundamental principles. At the same time, pedagogy is the broad principles and strategies of classroom management and organisation. During the process of preparing educators, it was observed that subject matter knowledge and teaching strategies were considered distinct components. This observation led Shulman in 1987 to introduce the concept of pedagogical content knowledge. This concept is recognised as\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e“the blending of content and pedagogy into an understanding of how particular topics, problems or issues are organised, represented, and adapted to the diverse interests and abilities of learners and presented for instruction”\u003c/em\u003e\u0026nbsp; Shulman (1987).\u003c/p\u003e\n\u003cp\u003eThis means that it is important for teachers to possess knowledge of what to teach and how to teach learners. After realising that the rapid changes in the technological landscape were impacting the learning-teaching process, Mishra and Koehler (2006) proposed that technological knowledge was critical to teachers’ knowledge, and thus, the TPCK was formulated. TPCK is described as\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e“… the basis of good teaching with technology and requires an understanding of the representation of concepts using technologies; pedagogical techniques that use technologies in constructive ways to teach content; knowledge of what makes concepts difficult or easy to learn and how technology can help redress some of the problems that students face”\u0026nbsp;\u003c/em\u003eMishra and Koehler (2006).\u003c/p\u003e\n\u003cp\u003eFrom the TPCK framework, physical sciences teachers require not just knowledge of what to teach and how to teach but also how the content they teach can be enhanced by using digital technology and the teaching strategies by which this can be achieved.\u003c/p\u003e"},{"header":"METHODOLOGY","content":"\u003cp\u003e\u003cstrong\u003eResearch Context\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe participants were part of a final year Bachelor of Education Honours degree class that took a module on the Methodology of Physical Sciences. In that module, one of the topics was integrating digital technology in teaching sciences, and ChatGPT was used as one of the emerging technologies. The instructor discussed the benefits and limitations of ChatGPT in educational contexts. The participants were instructed on designing prompts they could use in brainstorming in various situations, such as designing lesson plans and searching for appropriate learning/teaching resources such as multimedia, simulations, and virtual experiments. In this context, examples of prompts they would use were:\u003c/p\u003e\n\u003cp\u003eo\u0026nbsp;\u0026nbsp;Design a lesson plan utilising the 5E instructional approach to teach grade 10 learners the topic of electric circuits.\u003c/p\u003e\n\u003cp\u003eo\u0026nbsp;\u0026nbsp;Suggest a formative assessment task that can be used in teaching this topic.\u003c/p\u003e\n\u003cp\u003eo\u0026nbsp;\u0026nbsp;Suggest suitable teaching/learning resources such as simulations or virtual experiments.\u003c/p\u003e\n\u003cp\u003eWhile the instructor relied on traditional sources of information such as research articles and textbooks to expose the student teachers to the theories underpinning the methodologies of teaching physical sciences, the student teachers were encouraged to use ChatGPT to supplement their understanding. When they found useful information, they could adapt it to the curriculum they were being trained to teach. It was made clear that it is unethical to copy the responses from ChatGPT and present them as their own work. They had to adapt these responses to their own situations. This way, ChatGPT was used to supplement, enrich and complement the module\u0026apos;s teaching.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eSampling\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eParticipants were recruited using the purposive sampling technique to select participants who had shown greater interest and understanding of the integration of ChatGPT in science teaching. The participants were informed of the purpose of the study, and the researchers sought their informed consent, which they granted. The participants were also informed that they were free to withdraw from the study at any time without any adverse effects to them. They were also informed that their responses would be used for academic purposes only and that their identities would not be revealed to anyone. Their responses would remain anonymous. Eleven participants agreed to take part in the study. The data of the participants are summarised in Table 1. The names appearing in Table 1 are pseudonyms.\u003c/p\u003e\n\u003cp\u003eTable\u0026nbsp;1\u0026nbsp;Data of the Participants\u003c/p\u003e\n\u003cdiv align=\"Left\"\u003e\n \u003ctable border=\"0\" cellspacing=\"0\" cellpadding=\"0\"\u003e\n \u003ctbody\u003e\n \u003ctr\u003e\n \u003ctd width=\"57.142857142857146%\" valign=\"top\"\u003e\n \u003cp\u003eName of Participant\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.367346938775512%\" valign=\"top\"\u003e\n \u003cp\u003eGender\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"24.489795918367346%\" valign=\"top\"\u003e\n \u003cp\u003eAge\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"57.142857142857146%\" valign=\"top\"\u003e\n \u003cp\u003eJohn\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.367346938775512%\" valign=\"top\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"24.489795918367346%\" valign=\"top\"\u003e\n \u003cp\u003e26\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"57.142857142857146%\" valign=\"top\"\u003e\n \u003cp\u003eBennie\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.367346938775512%\" valign=\"top\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"24.489795918367346%\" valign=\"top\"\u003e\n \u003cp\u003e25\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"57.142857142857146%\" valign=\"top\"\u003e\n \u003cp\u003eMary\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.367346938775512%\" valign=\"top\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"24.489795918367346%\" valign=\"top\"\u003e\n \u003cp\u003e27\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"57.142857142857146%\" valign=\"top\"\u003e\n \u003cp\u003eElizabeth\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.367346938775512%\" valign=\"top\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"24.489795918367346%\" valign=\"top\"\u003e\n \u003cp\u003e25\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"57.142857142857146%\" valign=\"top\"\u003e\n \u003cp\u003eIrene\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.367346938775512%\" valign=\"top\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"24.489795918367346%\" valign=\"top\"\u003e\n \u003cp\u003e25\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"57.142857142857146%\" valign=\"top\"\u003e\n \u003cp\u003eEva\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.367346938775512%\" valign=\"top\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"24.489795918367346%\" valign=\"top\"\u003e\n \u003cp\u003e23\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"57.142857142857146%\" valign=\"top\"\u003e\n \u003cp\u003eAsante\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.367346938775512%\" valign=\"top\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"24.489795918367346%\" valign=\"top\"\u003e\n \u003cp\u003e26\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"57.142857142857146%\" valign=\"top\"\u003e\n \u003cp\u003eRosemary\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.367346938775512%\" valign=\"top\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"24.489795918367346%\" valign=\"top\"\u003e\n \u003cp\u003e30\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"57.142857142857146%\" valign=\"top\"\u003e\n \u003cp\u003ePeter\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.367346938775512%\" valign=\"top\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"24.489795918367346%\" valign=\"top\"\u003e\n \u003cp\u003e25\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"57.142857142857146%\" valign=\"top\"\u003e\n \u003cp\u003eRodrick\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.367346938775512%\" valign=\"top\"\u003e\n \u003cp\u003eM\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"24.489795918367346%\" valign=\"top\"\u003e\n \u003cp\u003e24\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003ctr\u003e\n \u003ctd width=\"57.142857142857146%\" valign=\"top\"\u003e\n \u003cp\u003eLebo\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"18.367346938775512%\" valign=\"top\"\u003e\n \u003cp\u003eF\u003c/p\u003e\n \u003c/td\u003e\n \u003ctd width=\"24.489795918367346%\" valign=\"top\"\u003e\n \u003cp\u003e42\u003c/p\u003e\n \u003c/td\u003e\n \u003c/tr\u003e\n \u003c/tbody\u003e\n \u003c/table\u003e\n\u003c/div\u003e\n\u003cp\u003e\u003cstrong\u003eData Collection\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eSemi-structured interviews were used in data collection. In conducting semi-structured interviews in qualitative research, the researcher uses open-ended questions to construct an interview guide (Magaldi \u0026amp; Berler, 2020). During the interview, the researcher uses probing follow-up questions to obtain an in-depth understanding of the phenomena from the participant\u0026rsquo;s viewpoint. This was the approach used in this study. The following interview guide was used to conduct the interviews:\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePart A: General Questions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eo\u0026nbsp;\u0026nbsp;What do you understand by AI, like ChatGPT?\u003c/p\u003e\n\u003cp\u003eo\u0026nbsp;\u0026nbsp;In your own words, what is ChatGPT?\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePart B: Use of ChartGPT in Lesson Preparation and in Learning and Teaching\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eo\u0026nbsp;How can ChatGPT be used in teaching physical sciences?\u003c/p\u003e\n\u003cp\u003eo\u0026nbsp;How would you use ChatGPT in the following:\u003c/p\u003e\n\u003cp\u003e\u0026middot;\u0026nbsp; \u0026nbsp;\u0026nbsp;Lesson planning, preparation and presentation\u003c/p\u003e\n\u003cp\u003e\u0026middot;\u0026nbsp; \u0026nbsp;\u0026nbsp;Assessment?\u003c/p\u003e\n\u003cp\u003e\u0026middot;\u0026nbsp; \u0026nbsp;\u0026nbsp;Ensuring learners produce prompts suitable to get useful responses from ChatGPT?\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePart C: Limitations of ChatGPT\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eo\u0026nbsp;ChatGPT can produce inaccurate information. How can you, as a teacher, deal with this?\u003c/p\u003e\n\u003cp\u003eo\u0026nbsp;What are some of the limitations of ChatGPT?\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePart D: Preservice Teachers\u0026rsquo; Perceptions of Integrating ChatGPT in Teaching Physical Sciences\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eo\u0026nbsp;Based on your understanding of the use of ChatGPT in teaching Physical sciences, how comfortable or prepared are you to integrate it into your teaching?\u003c/p\u003e\n\u003cp\u003eo\u0026nbsp;What areas in your learning of using ChatGPT in teaching require improvement?\u003c/p\u003e\n\u003cp\u003eo\u0026nbsp;What can be done to improve preservice teachers\u0026rsquo; use of ChatGPT in teaching physical sciences?\u003c/p\u003e\n\u003cp\u003eSemi-structured interviews were conducted using the above interview guide. An electronic device was used to record the participants\u0026rsquo; responses to the interview questions after obtaining their permission. Each interview lasted between 30 to 40 minutes. After the interviews, the participants\u0026rsquo; responses were transcribed verbatim using software. The participants were allowed to read their transcribed interviews. The participants agreed that the transcribed interviews represented their views. After authentification by member checking, the interview transcripts were ready for analysis.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDATA ANALYSIS\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eData analysis was done using computer-assisted qualitative data analysis software (CAQDAS), ATLAS.ti 8. We relied on Braun and Clarke (2006) six recursive stages of qualitative thematic analysis viz - getting familiar with the data; codes generation; identifying themes; reviewal of themes; defining and naming themes and report writing. The purpose of thematic analysis was to identify, analyse, report and interpret patterns within the data (Braun \u0026amp; Clarke, 2006). The transcribed interview transcripts were imported to ATLAS.ti 8, and the software was used to search for codes within the interview transcripts. Similar codes were then placed in groups to develop sub-themes and themes.\u003c/p\u003e"},{"header":"RESULTS AND DISCUSSION","content":"\u003cp\u003eThe themes that emerged from the data are used to organise the results of this study in this section. These results are discussed in the context of the literature reviewed.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePreservice Teachers\u0026rsquo; Understanding of Artificial Intelligence and ChatGPT\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eTwo sub-themes, respectively, represented the preservice teachers\u0026rsquo; understanding of AI and ChatGPT:\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eo\u0026nbsp;AI as machines and computers mimicking human intelligence\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eo\u0026nbsp;CahtGPT: AI based on GPT - a language model.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eThe preservice teachers understood AI technology as machines or computer programmes simulating human intelligence. This is what Bennie had to say when asked what he understood by AI:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026hellip; the theory and development of computer systems capable of performing human tasks or tasks requiring human intelligence,\u003c/em\u003e Bennie.\u003c/p\u003e\n\u003cp\u003eSimilar sentiments were echoed by Elizabeth and Irene, who described AI as follows:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026hellip;. involves technology that tries to simulate human intelligence by using machines,\u0026nbsp;\u003c/em\u003eElizabeth.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003e\u0026hellip;. the use of machines or computer systems to do what humans can or solve problems that can only be solved by a human,\u0026nbsp;\u003c/em\u003eIrene.\u003c/p\u003e\n\u003cp\u003eThe preservice teachers\u0026rsquo; understanding of ChatGPT as \u003cem\u003eAI based on GPT - a language model\u0026nbsp;\u003c/em\u003eis demonstrated by the following quotes:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eOkay, ChatGPT is a chatbot \u0026hellip; generative pre-trained transformer. Looking at what it stands for, I\u0026apos;ll say people train it; it\u0026apos;s trained by humans to converse with the user in a human-like form,\u0026nbsp;\u003c/em\u003eEva.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eI understand that it stands for Generative Pre-trained Transformer, and you give it prompts, and then it responds to you like a human being, but it\u0026apos;s a computer program,\u0026nbsp;\u003c/em\u003eLebo.\u003c/p\u003e\n\u003cp\u003eThe above quotes suggest that the preservice teachers understood artificial intelligence, ChatGPT, and their capabilities. Reiss (2021) claim that definitions of AI have evolved over time as new discoveries were made and see AI as an inanimate matter exhibiting some form of intelligence similar to human intelligence, as suggested by the preservice teachers. This includes machines such as robots performing tasks such as medical diagnosis, and software programs called bots mimicking human intelligence (Reiss, 2021). The understanding of the preservice teachers of ChatGPT as a pre-trained transformer, or a natural language model, trained on a huge amount of words (billions) to be able to respond to text inputs is closely aligned \u0026nbsp;with literature (Sharma \u0026amp; Yadav, 2022).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLesson Preparation, Planning and Presentation\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe preservice physical science teachers believed that ChatGPT could be helpful in lesson preparation, planning and presentation. Under this theme, their responses were placed in the following sub-themes: lesson objectives, official curriculum documents, lesson preparation, planning, presentation, summarising content, ChatGPT as a resource for teaching/learning materials, simulations, ChatGPT as a tool for learner engagement and ChatGPT as a tool for assessment. Figure 1 demonstrates the relationships between these sub-themes.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eRegarding lesson objectives, participant Eva had this to say:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eWhen ChatGPT gives you the objectives, you, as a teacher, must check them and critically evaluate them. Then, after you evaluate them, you go to a CAPS document or ATP, and then you combine them so that you can provide or come up with the proper lesson plan,\u003c/em\u003e Eva.\u003c/p\u003e\n\u003cp\u003eParticipants were of the view that curriculum documents must be consulted in lesson preparation, planning, and presentation. As mentioned by Eva, the objectives in any topic that the teacher is teaching while integrating ChatGPT must be verified against curriculum documents. In the case of this group of students, the curriculum documents were the Curriculum and Assessment Policy Statement (CAPS), the Annual Teaching Plans (ATPs) and examination guidelines. Regarding \u003cem\u003eofficial curriculum documents,\u003c/em\u003e Asante had this to say:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAs a teacher, I will always use other sources to verify my information, to check if the kind of information I am getting from ChatGPT is appropriate or the kind of information I want. Not just copying everything from ChatGPT. I will use the CAPS document and the different textbooks to check if it\u0026rsquo;s the correct information or if it\u0026apos;s wrong,\u0026nbsp;\u003c/em\u003eAsante.\u003c/p\u003e\n\u003cp\u003eIn terms of \u003cem\u003elesson preparation\u003c/em\u003e, this is what Lebo had to say:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eFor lesson preparation, I think mostly to assist me with teaching aids because then it will direct me as to where I can find the simulations for the hands-on experiments and also give me ideas on what type of experiments I can perform with learners for them to understand that particular concept that I want to teach about\u003c/em\u003e, Lebo.\u003c/p\u003e\n\u003cp\u003eOther participants, just like Lebo, mentioned that ChatGPT was useful in searching for resources such as simulations and videos, websites and summarising content. Under the sub-theme of \u003cem\u003eassessment in physical sciences\u003c/em\u003e, Mary said:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eIt can help teachers who teach physical sciences know how to use ChatGPT to assess learners or create questions they can use in their lessons. In terms of assessment, I can use it to generate questions, maybe like multiple choice, true or false questions, long and short questions,\u0026nbsp;\u003c/em\u003eMary.\u003c/p\u003e\n\u003cp\u003eUnder the sub-theme of \u003cem\u003eChatGPT as a tool for learner engagement\u003c/em\u003e, Irene said:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eChatGPT can help engage the learners. The students learn better when they\u0026apos;re interacting in class. If I can give a question to learners and ask them to please check this information from ChatGPT, it will help learners be active and engaged in the classroom\u003c/em\u003e, Irene.\u003c/p\u003e\n\u003cp\u003eOther participants expressed similar sentiments, indicating that preservice teachers regarded ChatGPT as a useful tool for enhancing learner engagement and generating assessment tasks and teaching/learning resources. The sentiments expressed by the preservice teachers on the potential benefits of integrating ChatGPT in teaching physical sciences align with findings from previous studies. For example, a study by Jo (2024)\u0026nbsp; revealed that using ChatGPT in learning results in enhanced knowledge acquisition and application, with the novelty associated with personalised learning using ChatGPT stimulating engagement, thus making learning exciting and motivating. The preservice teachers\u0026apos; perceptions that ChatGPT is a useful tool for generating questions or responding to learners\u0026rsquo; queries align with the findings from the literature. A study by Lee and Zhai (2024) revealed that preservice science teachers regarded ChatGPT as useful in learners\u0026apos; formative assessment. This is supported by Rahman and Watanobe (2023) who claim that physics teachers can use ChatGPT to generate practice questions, quizzes and explanations for formative assessment. They also argue that ChatGPT can provide step-by-step solutions, thereby developing learners\u0026rsquo; problem-solving skills. Rahman and Watanobe (2023) also reported that ChatGPT is a useful tool in lesson planning in chemistry and physics, as indicated by the preservice teachers. This statement corresponds with the research findings indicating that ChatGPT is capable of creating effective lesson plans based on the 5E model (Cooper, 2023).\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003ePrompt Engineering\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe participants\u0026apos; responses on prompt engineering reflect that they had an appreciable level of understanding. When asked to describe how they could assist their learners to ask prompts in ChatGPT, some of their responses were as follows:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eBecause, you see, your students are in grade 10, grade 11 or 12, and the prompts they give must be specific for that level of education. Because ChatGPT can sometimes provide information that is too advanced for the students. So, on the ChatGPT, you need to ask a specific question related to that grade or that standard so that it can provide you with the correct information. And you need to ask for specific information so that it won\u0026apos;t give you the wrong information or the more advanced information to that standard,\u0026nbsp;\u003c/em\u003eRosemary.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eLet\u0026apos;s say we are doing electrical circuits, and they have a question about how resistors in parallel affect the current. They must also be specific that we are doing it for grade 10. The content needs to be for a grade 10 learner so that then when it gives out the answers, the answers will be in such a way that it\u0026apos;s understandable to a learner that is in grade 10 rather than maybe an electrician who then has more knowledge about electricity than a grade 10 learner,\u0026nbsp;\u003c/em\u003eLebo.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eI can ask my learners to ask clear, specific questions. If possible, they must also provide context and follow-up questions,\u0026nbsp;\u003c/em\u003ePeter.\u003c/p\u003e\n\u003cp\u003eThe responses of the participants echo some of the features of prompt engineering that are discussed in Giray (2023), who defines a prompt as \u003cem\u003especific instruction\u003c/em\u003e or a \u003cem\u003equery\u003c/em\u003e that a learner provides to ChatGPT. The four elements of effective prompting identified by Giray (2023) are: instruction, context, input data and output indicator. As noted by preservice teachers, the instruction has to be specific to guide the chatbot to an appropriate response. The context is the additional information that the learner provides in the prompt to guide ChatGPT to generate an accurate output (Giray, 2023). Peter stated that learners have to be taught to provide context, and other participants, such as Rosemary, stated that learners have to be specific and that they need responses suitable for the grade 10 level of education, for example.\u003c/p\u003e\n\u003cp\u003eRegarding input data, Giray (2023) describes this as the core of the prompt and is the specific question that ChatGPT should respond to. This is discussed by Lebo and Rodrick as follows:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eI think I will have to train or assist my learners in properly putting the prompts so that they give the relevant information. Because if you don\u0026apos;t give it the correct prompts, it will just give you any information, seeing that it has access to a lot of information. So, learners need to know that there are certain words that they need to use that will assist them in getting specific answers from ChatGPT\u003c/em\u003e, Lebo.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eWe must be able to teach our learners how to prompt ChatGPT so that they can get problems, explanations, and summaries on topics we\u0026apos;ve taught them\u003c/em\u003e, Rodrick.\u003c/p\u003e\n\u003cp\u003eAs for the output indicator, the learner must direct ChatGPT regarding the response format required (Giray, 2023). The learner can guide ChatGPT about the length of the response and the details of the response sought. While Rodrick stated the need to specify the type of response, such as problems, summary or explanation, the participant fell short of stating that learners must describe the length of the response.\u0026nbsp;\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eLimitations of ChatGPT and Suggestions on Mitigating the Limitations\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe major limitation that ChatGPT currently possesses, which was of great concern to preservice physical sciences teachers, was that sometimes ChatGPT provided inaccurate information (Elmas et al., 2024). While the teachers claimed they could notice inaccurate information, they feared that their learners would be unable to differentiate between accurate and inaccurate information. Some of the quotes from the participants are as follows:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAs ChatGPT can provide inaccurate information, as a teacher, I will always use other sources to verify my information and check if the kind of information I am getting from ChatGPT is appropriate or the kind of information I want. Not just copying everything from ChatGPT. I will use the CAPS document and the different textbooks to check if that\u0026apos;s the correct information or if it\u0026apos;s wrong\u003c/em\u003e, Asante.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eWhen I\u0026apos;m teaching my learners, when I get the information from ChatGPT, I have to go through it, and I have to consult some sources before I go and teach my learners such information, and I have to use it as an aid\u003c/em\u003e, Bennie.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eThe teacher must emphasise the limitations of using ChatGPT, as many people or learners don\u0026apos;t know where they have to stop. They take all the information, so they must have the knowledge of the limitations. We need to know a lot about limitations\u003c/em\u003e, Elizabeth.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eI can always teach students that they have to learn a good way of putting in the prompt questions, and they don\u0026apos;t have to rely too much on it. They have to take it as an assistant\u003c/em\u003e, John.\u003c/p\u003e\n\u003cp\u003eThe teachers were of the view that using ChatGPT in conjunction with traditional sources of information would mitigate some of its constraints. However, while they claimed that they would do this, it was not clear if learners would also manage to utilize ChatGPT in this manner as refelected by the following quotes:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eAs a teacher, on my side, I think I have an advantage because I already have the content knowledge. So, I\u0026apos;ll be able to identify that this information is incorrect and then use other resources, be it a textbook or other resources, to get the correct information\u003c/em\u003e. \u003cem\u003eFor me, it starts being dangerous on the part of the learner because the learner is not equipped enough to realise that this information is incorrect. And looking at our children these days, they are over-reliant on digital information. So that\u0026apos;s where it will be a challenge as a teacher to encourage my learners to get proper content; they need not rely on ChatGPT only. They also need to refer to the proper textbook we use for physical sciences to get the correct information,\u0026nbsp;\u003c/em\u003eLebo.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eA teacher should understand the content so that we can see the flaws in the answers that are out there. Sometimes, it\u0026apos;s not just ChatGPT; even the learners\u0026apos; guides sometimes have the wrong information. So, a teacher should be able to understand the content. ChatGPT, though it\u0026apos;s worse than the guides that the learners sometimes use, requires a teacher to understand the subject matter and the subject content. A teacher must understand the curriculum document. A teacher must understand the exam guidelines so that we can see the flaws in the answers from ChatGPT or explanations\u003c/em\u003e, Rodrick.\u003c/p\u003e\n\u003cp\u003eThe constraints identified by the preservice educators, such as the presence of inaccurate responses, are likely a result of the existing capabilities of chatbots. A study by Fergus et al. (2023) revealed that in chemistry, ChatGPT can satisfactorily provide answers to questions at the remembering and comprehension levels in Bloom\u0026rsquo;s taxonomy and struggles with questions that require application or interpretation. They particularly noted that the chatbot faces difficulties with nontext information, such as questions with graphical or numeric information.\u0026nbsp;Additionally, comparable findings showed that ChatGPT was unable to achieve over 37% accuracy in responses to multiple-choice questions or brief answers within an introductory-level university chemistry course (Leon \u0026amp; Vidhani, 2023). This means that learners would get incorrect responses most of the time when they attempted to use ChatGPT. Poor performance in responding to biology questions was also observed by Elmas et al. (2024).\u003c/p\u003e\n\u003cp\u003eAnother potential limitation of ChatGPT identified by preservice teachers was that learners could develop over-reliance on it. There was fear that ChatGPT could assume the role of an epistemic authority for learners despite the limitation of providing inaccurate information in some chemistry and physics problems. Some of the quotes illustrating this are:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eBut it\u0026apos;s\u0026nbsp;\u003c/em\u003e(ChatGPT) \u003cem\u003ealso not good at all for our learners. As we know, learners are much lazier when searching for information. They will rely much more on ChatGPT as we show them how to use it,\u0026nbsp;\u003c/em\u003eRosemary.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eThe people\u0026nbsp;\u003c/em\u003e(learners)\u003cem\u003e\u0026nbsp;using it need to use it responsibly. Science is not just about getting a correct answer but how you get to that answer. So, in terms of developing critical skills and problem-solving, that still needs to be done by a human being. So, people need not over-rely on it to provide the correct answer and not understand how to get to that correct answer\u003c/em\u003e, Lebo.\u003c/p\u003e\n\u003cp\u003ePreservice teachers believed that while ChatGPT could be leveraged to provide science information, traditional sources of information should be used to verify that information. They noted that this would be difficult for learners.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eYes, and I\u0026apos;ll also advise them that when they get information on ChatGPT, they must not take it as it is, but they must look for another source of information and combine it with that information from ChatGPT\u003c/em\u003e, Rosemary.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eFor me, it starts being dangerous on the part of the learner because the learner is not equipped enough to realise that this information is incorrect. And looking at our children these days, they are over-reliant on digital information. So that\u0026apos;s where it will be a challenge as a teacher to encourage my learners to get proper content; they need not rely on ChatGPT only. They also need to refer to the proper textbook we use for physical sciences to get the correct information\u003c/em\u003e, Lebo.\u003c/p\u003e\n\u003cp\u003eThese results are aligned with the literature which suggests that developers of ChatGPT should make an effort to ensure that their algorithms produce more accurate responses in education to prevent harm to learners \u0026nbsp;(Cambra-Fierro et al., 2024). While empirical studies indicate that ChatGPT has many potential benefits at different levels of education ranging from increasing the teachers\u0026rsquo; well-being and reducing stress and assisting in the achievement of learning objectives and creation of effective lesson plans (Cambra-Fierro et al., 2024; Wijaya et al., 2024), there is a serious need for teachers to be aware of the limitations discussed by the preservice teachers.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAreas of Improvement\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe preservice teachers identified auto-marking learners\u0026rsquo; works, preventing plagiarism and how teachers can be assisted in dealing with inaccurate responses as areas that needed improvement in preparing preservice teachers to integrate ChatGPT in teaching Physical Sciences. The following quotes exemplify their sentiments:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eWe are not quite sure about the use of ChatGPT to assess learners\u0026apos; work.\u0026nbsp;\u003c/em\u003e\u003cem\u003eI think they are not being taught enough. So the teachers do not emphasise them. Limitations.\u0026nbsp;\u003c/em\u003e\u003cem\u003eI think it\u0026apos;s the same as the limitations. But it has got limitations. Such as providing incorrect responses. And I\u0026rsquo;m unsure how to deal with that in your teaching practice\u003c/em\u003e, Mary.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eThe area that must be improved is how learners can avoid plagiarism from ChatGPT. In higher education, most assessments are done online, and learners use ChatGPT as a source of plagiarism.\u0026nbsp;\u003c/em\u003e\u003cem\u003eI think they need to come up with solutions I\u0026apos;m not sure if there can be apps that can be used to help to avoid plagiarism\u003c/em\u003e, Peter.\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eThe area that I did not find well was that when writing assignments, ChatGPT makes it too easy for a person to write an assignment; you may end up not learning that much because you rely on ChatGPT; you only need a summary, and then you can go to ChatGPT to have an idea after having an idea of what you are searching on ChatGPT can assist you. Therefore, it will limit the critical thinking of students\u003c/em\u003e, Irene.\u003c/p\u003e\n\u003cp\u003eThe participants\u0026apos; responses imply that preservice teachers must develop their skills in automatic grading of student work and explore how well they can incorporate these techniques into their teaching practices. The physical sciences preservice teachers indicated that they were very keen to incorporate ChatGPT in their teaching practices, as expressed by Rodrick in the following quote:\u003c/p\u003e\n\u003cp\u003e\u003cem\u003eJust incorporating more on the use of AI. We are in the 4IR at this time, and technology is taking over. So, we cannot be left behind as physical science teachers. Upcoming physical science teachers must be taught how to use AI in their teaching. How can they use technology in their teaching?\u003c/em\u003e Rodrick.\u003c/p\u003e\n\u003cp\u003eThe preservice teachers\u0026rsquo; perceptions regarding the benefits of integrating ChatGPT in learning physical sciences, such as explaining complex science concepts, assisting in problem-solving and conceptual understanding, and increasing the learners\u0026rsquo; motivation, interest and engagement, are supported by literature (Iyamuremye \u0026amp; Ndihokubwayo, 2024; Kodkin \u0026amp; Artem\u0026rsquo;eva, 2024; Taani \u0026amp; Alabidi, 2024). However, it is important that teachers should take cognisance of the limitations of ChatGPT, particularly in responding to questions that require application and solving complex problems in physics and chemistry (Leon \u0026amp; Vidhani, 2023).\u003c/p\u003e"},{"header":"Conclusion","content":"\u003cp\u003eThe study has provided empirical evidence for the potential areas in which ChatGPT can be used to enhance physical sciences instruction. These areas include lesson preparation, planning and presentation. In lesson preparation and planning, the preservice teachers believed that ChatGPT is a tool that can be used for content generation, searching for teaching/learning materials, and generating formative assessment tasks. ChatGPT was also found to have the potential to assist in generating innovative lesson plans. Nevertheless, the investigation found that ChatGPT should function in a supplementary capacity. It is advisable to use it to complement conventional sources of information, such as textbooks and formal curriculum guides. It is advisable for the creators of ChatGPT to enhance the accuracy of the generated responses to questions about the physical sciences. Currently, the application often provides incorrect answers, which may contribute to misconceptions in fields such as physics and chemistry. This could adversely affect the education in these areas. Policymakers should ensure that they provide professional development to practising teachers on the potential benefits and constraints of ChatGPT in physical sciences instruction.\u003c/p\u003e\n\u003cp\u003eFurther investigation is advised to assess ChatGPT's influence on physical sciences education. Implementing experimental and quasi-experimental research in educational settings could mitigate some of the limitations identified in this research, which was constrained by its case study approach and the small sample size from a rural area.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors would like to express their gratitude to the participants for taking the time to respond to the interviews and to the university research ethics committee for granting them permission to conduct the study.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eDeclaration of Interest\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no conflict of interest.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eBaidoo-Anu, D., \u0026amp; Ansah, L. O. (2023). Education in the era of generative artificial intelligence (AI): Understanding the potential benefits of ChatGPT in promoting teaching and learning. \u003cem\u003eJournal of AI\u003c/em\u003e,\u003cem\u003e 7\u003c/em\u003e(1), 52-62. https://doi.org/10.61969/jai.1337500\u003c/li\u003e\n\u003cli\u003eBarak, M. (2017). Science teacher education in the twenty-first century: A pedagogical framework for technology-integrated social constructivism. \u003cem\u003eResearch in Science Education\u003c/em\u003e,\u003cem\u003e 47\u003c/em\u003e, 283-303. https://doi.org/10.1007/s11165-015-9501-y\u003c/li\u003e\n\u003cli\u003eBraun, V., \u0026amp; Clarke, V. (2006). Using thematic analysis in psychology. \u003cem\u003eQualitative research in psychology\u003c/em\u003e,\u003cem\u003e 3\u003c/em\u003e(2), 77-101. https://doi.org/10.1191/1478088706qp063oa\u003c/li\u003e\n\u003cli\u003eCambra-Fierro, J. J., Blasco, M. F., L\u0026oacute;pez-P\u0026eacute;rez, M.-E. E., \u0026amp; Trifu, A. (2024). ChatGPT adoption and its influence on faculty well-being: An empirical research in higher education. \u003cem\u003eEducation and Information Technologies\u003c/em\u003e. https://doi.org/10.1007/s10639-024-12871-0\u003c/li\u003e\n\u003cli\u003eChiu, T. K., Moorhouse, B. L., Chai, C. S., \u0026amp; Ismailov, M. (2023). Teacher support and student motivation to learn with Artificial Intelligence (AI) based chatbot. \u003cem\u003eInteractive Learning Environments\u003c/em\u003e, 1-17. https://doi.org/10.1080/10494820.2023.2172044\u003c/li\u003e\n\u003cli\u003eCooper, G. (2023). Examining science education in ChatGPT: An exploratory study of generative artificial intelligence. \u003cem\u003eJournal of Science Education and Technology\u003c/em\u003e,\u003cem\u003e 32\u003c/em\u003e(3), 444-452. https://doi.org/10.1007/s10956-023-10039-y\u003c/li\u003e\n\u003cli\u003eDeng, J. M., Lalani, Z., McDermaid, L. A., \u0026amp; Szozda, A. R. (2023). Using generative artificial intelligence in chemistry education research: prioritizing ethical use and accessibility. \u003cem\u003eChamical Education\u003c/em\u003e. https://doi.org/10.26434/chemrxiv-2023-24zfl\u003c/li\u003e\n\u003cli\u003eDervić, D., Glamočić, D. S., Gazibegović-Busuladžić, A., \u0026amp; Me\u0026scaron;ić, V. (2018). Teaching physics with simulations: Teacher-centered versus student-centered approaches. \u003cem\u003eJournal of Baltic Science Education\u003c/em\u003e,\u003cem\u003e 17\u003c/em\u003e(2), 288. https://doi.org/10.33225/jbse/18.17.288\u003c/li\u003e\n\u003cli\u003eElmas, R., Adiguzel-Ulutas, M., \u0026amp; Yılmaz, M. (2024). Examining ChatGPT\u0026rsquo;s validity as a source for scientific inquiry and its misconceptions regarding cell energy metabolism. \u003cem\u003eEducation and Information Technologies\u003c/em\u003e. https://doi.org/10.1007/s10639-024-12749-1\u003c/li\u003e\n\u003cli\u003eElsen-Rooney, M. (2023). \u003cem\u003eNYC education department blocks ChatGPT on school devices, networks\u003c/em\u003e. Chalkbeat New York. Retrieved 20 April from https://www.chalkbeat.org/newyork/2023/1/3/23537987/nyc-schools-ban-chatgpt-writing-artificial-intelligence/\u003c/li\u003e\n\u003cli\u003eFergus, S., Botha, M., \u0026amp; Ostovar, M. (2023). Evaluating academic answers generated using ChatGPT. \u003cem\u003eJournal of chemical Education\u003c/em\u003e,\u003cem\u003e 100\u003c/em\u003e(4), 1672-1675. https://doi.org/10.1021/acs.jchemed.3c00087\u003c/li\u003e\n\u003cli\u003eGill, S. S., Xu, M., Patros, P., Wu, H., Kaur, R., Kaur, K., Fuller, S., Singh, M., Arora, P., \u0026amp; Parlikad, A. K. (2024). Transformative effects of ChatGPT on modern education: Emerging Era of AI Chatbots. \u003cem\u003eInternet of Things and Cyber-Physical Systems\u003c/em\u003e,\u003cem\u003e 4\u003c/em\u003e, 19-23. https://doi.org/10.1016/j.iotcps.2023.06.002\u003c/li\u003e\n\u003cli\u003eGiray, L. (2023). Prompt engineering with ChatGPT: a guide for academic writers. \u003cem\u003eAnnals of biomedical engineering\u003c/em\u003e,\u003cem\u003e 51\u003c/em\u003e(12), 2629-2633. https://doi.org/10.1007/s10439-023-03272-4\u003c/li\u003e\n\u003cli\u003eGrassini, S. (2023). Shaping the future of education: exploring the potential and consequences of AI and ChatGPT in educational settings. \u003cem\u003eEducation Sciences\u003c/em\u003e,\u003cem\u003e 13\u003c/em\u003e(7), 692. https://doi.org/10.3390/educsci13070692\u003c/li\u003e\n\u003cli\u003eIyamuremye, A., \u0026amp; Ndihokubwayo, K. (2024). Exploring secondary school students\u0026rsquo; interest and mastery of atomic structure and chemical bonding through ChatGPT. \u003cem\u003eEducational Journal of Artificial Intelligence and Machine Learning\u003c/em\u003e,\u003cem\u003e 1\u003c/em\u003e(1), 1-13. https://doi.org/10.58197/prbl/9hk37296\u003c/li\u003e\n\u003cli\u003eJo, H. (2024). From concerns to benefits: a comprehensive study of ChatGPT usage in education. \u003cem\u003eInternational Journal of Educational Technology in Higher Education\u003c/em\u003e,\u003cem\u003e 21\u003c/em\u003e(1), 35. https://doi.org/10.1186/s41239-024-00471-4\u003c/li\u003e\n\u003cli\u003eKodkin, V. L., \u0026amp; Artem\u0026rsquo;eva, E. V. (2024). ChatGPT: Application in Chemistry Education and Challenges. \u003cem\u003eJournal of Computer and Communications\u003c/em\u003e,\u003cem\u003e 12\u003c/em\u003e(03), 196-206. https://doi.org/10.4236/jcc.2024.123012\u003c/li\u003e\n\u003cli\u003eLee, G. G., \u0026amp; Zhai, X. (2024). Using ChatGPT for Science Learning: A Study on Pre-service Teachers\u0026apos; Lesson Planning. \u003cem\u003eIEEE Transactions on Learning Technologies\u003c/em\u003e. https://doi.org/10.1109/tlt.2024.3401457\u003c/li\u003e\n\u003cli\u003eLee, H. (2023). The rise of ChatGPT: Exploring its potential in medical education. \u003cem\u003eAnatomical sciences education\u003c/em\u003e. https://doi.org/10.1002/ase.2270\u003c/li\u003e\n\u003cli\u003eLeon, A. J., \u0026amp; Vidhani, D. (2023). ChatGPT needs a chemistry tutor too. \u003cem\u003eJournal of chemical Education\u003c/em\u003e,\u003cem\u003e 100\u003c/em\u003e(10), 3859-3865. https://doi.org/10.1021/acs.jchemed.3c00288\u003c/li\u003e\n\u003cli\u003eLin, X. (2023). Exploring the role of ChatGPT as a facilitator for motivating self-directed learning among adult learners. \u003cem\u003eAdult Learning\u003c/em\u003e. https://doi.org/10.1177/10451595231184928\u003c/li\u003e\n\u003cli\u003eLiu, Z., Vobolevich, A., \u0026amp; Oparin, A. (2023). The Influence of AI ChatGPT on Improving Teachers\u0026apos; Creative Thinking. \u003cem\u003eInternational Journal of Learning, Teaching and Educational Research\u003c/em\u003e,\u003cem\u003e 22\u003c/em\u003e(12), 124-139. https://doi.org/10.26803/ijlter.22.12.7\u003c/li\u003e\n\u003cli\u003eMagaldi, D., \u0026amp; Berler, M. (2020). Semi-structured interviews. \u003cem\u003eEncyclopedia of Personality and Individual Differences\u003c/em\u003e, 4825-4830. https://doi.org/10.1007/978-3-319-24612-3_857\u003c/li\u003e\n\u003cli\u003eMishra, P., \u0026amp; Koehler, M. J. (2006). Technological pedagogical content knowledge: A framework for teacher knowledge. \u003cem\u003eTeachers College Record\u003c/em\u003e,\u003cem\u003e 108\u003c/em\u003e(6), 1017-1054. https://doi.org/10.1111/j.1467-9620.2006.00684.x\u003c/li\u003e\n\u003cli\u003eNguyen, G., Nguyen, N., \u0026amp; Giang, N. (2022). Situation and Proposals for Implementing Artificial Intelligence-based Instructional Technology in Vietnamese Secondary Schools. \u003cem\u003eInternational Journal of Emerging Technologies in Learning\u003c/em\u003e,\u003cem\u003e 17\u003c/em\u003e(18), 53-75. https://doi.org/10.3991/ijet.v17i18.31503\u003c/li\u003e\n\u003cli\u003eRahman, M. M., \u0026amp; Watanobe, Y. (2023). ChatGPT for education and research: Opportunities, threats, and strategies. \u003cem\u003eApplied Sciences\u003c/em\u003e,\u003cem\u003e 13\u003c/em\u003e(9). https://doi.org/10.3390/app13095783\u003c/li\u003e\n\u003cli\u003eReiss, M. J. (2021). The Use of Al in Education: Practicalities and Ethical Considerations. \u003cem\u003eLondon Review of Education\u003c/em\u003e,\u003cem\u003e 19\u003c/em\u003e(1), n1. https://doi.org/10.14324/lre.19.1.05\u003c/li\u003e\n\u003cli\u003eSharma, S., \u0026amp; Yadav, R. (2022). Chat GPT\u0026ndash;A technological remedy or challenge for education system. \u003cem\u003eGlobal Journal of Enterprise Information System\u003c/em\u003e,\u003cem\u003e 14\u003c/em\u003e(4), 46-51. https://doi.org/10.18311/gjeis/2022\u003c/li\u003e\n\u003cli\u003eShulman, L. (1987). Knowledge and teaching: Foundations of the new reform. \u003cem\u003eHarvard Educational Review\u003c/em\u003e,\u003cem\u003e 57\u003c/em\u003e(1), 1-23. https://doi.org/10.17763/haer.57.1.j463w79r56455411\u003c/li\u003e\n\u003cli\u003eTaani, O., \u0026amp; Alabidi, S. (2024). ChatGPT in education: benefits and challenges of ChatGPT for mathematics and science teaching practices. \u003cem\u003eInternational Journal of Mathematical Education in Science and Technology\u003c/em\u003e, 1-30. https://doi.org/10.1080/0020739x.2024.2357341\u003c/li\u003e\n\u003cli\u003eWark, N., \u0026amp; Ally, M. (2020). An Emergent Pedagogical Framework for Integrating Emergent Technologies. In \u003cem\u003eEmerging Technologies and Pedagogies in the Curriculum\u003c/em\u003e, 89. https://doi.org/10.1007/978-981-15-0618-5_6\u003c/li\u003e\n\u003cli\u003eWest, J. K., Franz, J. L., Hein, S. M., Leverentz-Culp, H. R., Mauser, J. F., Ruff, E. F., \u0026amp; Zemke, J. M. (2023). An analysis of ai-generated laboratory reports across the chemistry curriculum and student perceptions of chatgpt. \u003cem\u003eJournal of chemical Education\u003c/em\u003e,\u003cem\u003e 100\u003c/em\u003e(11), 4351-4359. https://doi.org/10.1021/acs.jchemed.3c00581\u003c/li\u003e\n\u003cli\u003eWijaya, T. T., Su, M., Cao, Y., Weinhandl, R., \u0026amp; Houghton, T. (2024). Examining Chinese preservice mathematics teachers\u0026rsquo; adoption of AI chatbots for learning: Unpacking perspectives through the UTAUT2 model. \u003cem\u003eEducation and Information Technologies\u003c/em\u003e. https://doi.org/10.1007/s10639-024-12837-2\u003c/li\u003e\n\u003cli\u003eWollny, S., Schneider, J., Di Mitri, D., Weidlich, J., Rittberger, M., \u0026amp; Drachsler, H. (2021). Are we there yet?-a systematic literature review on chatbots in education. \u003cem\u003eFrontiers in Artificial Intelligence\u003c/em\u003e,\u003cem\u003e 4\u003c/em\u003e. https://doi.org/10.3389/frai.2021.654924\u003c/li\u003e\n\u003cli\u003eYavuz, A. C. (2020). The effects of differentiated instruction on Turkish students\u0026rsquo; L2 achievement, and student and teacher perceptions. \u003cem\u003eEurasian Journal of Applied Linguistics\u003c/em\u003e,\u003cem\u003e 6\u003c/em\u003e(2), 313-335. https://doi.org/10.32601/ejal.776002\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":true,"highlight":"","institution":"University of Venda","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true},"keywords":"Artificial intelligence, ChatGPT, physical sciences, technological, pedagogical content knowledge","lastPublishedDoi":"10.21203/rs.3.rs-4783356/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-4783356/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eThe emergence of artificial intelligence, exemplified by generative chatbots like ChatGPT, has elicited optimism among some educators regarding enhanced teaching and learning methods. Simultaneously, it has raised concerns among others who perceive these chatbots as disruptive to established pedagogical norms developed over centuries. This study investigated and analysed pre-service teachers' perceptions regarding integrating ChatGPT into teaching physical sciences at a rural university. A case study research design that used a qualitative approach was used to collect, analyse and interpret data. This methodology was employed to gain a comprehensive insight into the viewpoints held by physical science pre-service teachers. The study explored the benefits and potential challenges of incorporating emerging technologies like ChatGPT into teaching physical sciences. The theoretical framework that guided the study was the technological, pedagogical content knowledge. Eleven purposively sampled pre-service physical science teachers participated in semi-structured interviews. The collected data were analysed using thematic analysis. The research findings were that ChatGPT has the potential to contribute to teaching physical sciences in lesson planning, preparation, presentation and formative assessment. However, the study revealed that the inability of ChatGPT to answer some questions in physical sciences was of great concern. These findings shed light on how artificial intelligence generative chatbots can be incorporated into science teaching and learning. The findings provide insights for policymakers, science educators and researchers to deepen their understanding of the role of emerging technologies in science education.\u003c/p\u003e","manuscriptTitle":"Pre-service Physical Sciences Teachers’ Views on Integrating ChatGPT into Teaching: A Case Study","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2024-07-23 12:53:52","doi":"10.21203/rs.3.rs-4783356/v1","editorialEvents":[{"type":"communityComments","content":0}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"researchsquare","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":true,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"/submission","title":"Research Square","twitterHandle":"researchsquare","acdcEnabled":true,"dfaEnabled":false,"editorialSystem":"","reportingPortfolio":"","inReviewEnabled":false,"inReviewRevisionsEnabled":true}}],"origin":"","ownerIdentity":"d5ee0a06-4975-4eed-a63a-31b36a3b3173","owner":[],"postedDate":"July 23rd, 2024","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"posted","subjectAreas":[{"id":34966548,"name":"Artificial Intelligence and Machine Learning"}],"tags":[],"updatedAt":"2024-07-23T12:53:52+00:00","versionOfRecord":[],"versionCreatedAt":"2024-07-23 12:53:52","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-4783356","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-4783356","identity":"rs-4783356","version":["v1"]},"buildId":"qtupq5eGEP_6zYnWcrvyt","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}