Safety of Vision-Language-Action Models: A Survey from Lifecycle Perspectives

Safety of Vision-Language-Action Models: A Survey from Lifecycle Perspectives | Authorea try { document.documentElement.classList.add('js'); } catch (e) { } var _gaq = _gaq || []; _gaq.push(['_setAccount', 'G-8VDV14Y67G']); _gaq.push(['_trackPageview']); (function() { var ga = document.createElement('script'); ga.type = 'text/javascript'; ga.async = true; ga.src = ('https:' == document.location.protocol ? 'https://ssl' : 'http://www') + '.google-analytics.com/ga.js'; var s = document.getElementsByTagName('script')[0]; s.parentNode.insertBefore(ga, s); })(); Skip to main content Preprints Collections Wiley Open Research IET Open Research Ecological Society of Japan All Collections About About Authorea FAQs Contact Us Quick Search anywhere Search for preprint articles, keywords, etc. Search Search ADVANCED SEARCH SCROLL This is a preprint and has not been peer reviewed. Data may be preliminary. 3 April 2026 V1 Latest version Share on Safety of Vision-Language-Action Models: A Survey from Lifecycle Perspectives Authors : Wei Yuan 0000-0002-9400-842X [email protected] , Fengwen Liu , Ruize Wei , Zongwei Wang , Yuan Gao , Binxing Fang , Hu Huang , and Qing Liao Authors Info & Affiliations https://doi.org/10.22541/au.177524426.60806944/v1 435 views 189 downloads Contents Abstract Supplementary Material Information & Authors Metrics & Citations View Options References Figures Tables Media Share Abstract Vision-Language-Action (VLA) models have recently emerged as a promising paradigm for embodied intelligence, enabling robots to perform complex actions over multimodal observations with one end-to-end policy. With the increasing deployment of VLA models in real-world environments, ensuring their safety has become critical, as failures or malicious behaviors can result in severe physical harm and far-reaching societal consequences. Despite the growing body of research on VLA safety, existing studies largely focus on isolated aspects, such as specific attack surfaces, robustness, or reliability, without providing a unified understanding of safety risks across the full lifecycle of VLA systems. To bridge this gap, we present a comprehensive survey of safety challenges and mitigation strategies for VLA models from a holistic lifecycle perspective. Specifically, we decompose the VLA development pipeline into three key stages: data preparation, model training, and system deployment. For each stage, we systematically examine safety issues from two complementary perspectives: adversarial scenarios, where intentional attacks aim to compromise system safety, and non-adversarial settings, where inherent limitations and external factors degrade system reliability, leading to unsafe behaviors. Finally, we highlight emerging research directions toward building safe VLA systems. To support both researchers and practitioners in advancing the safe use of VLA systems, we also maintain an up-to-date repository of VLA safety literature at https://github.com/hi-weiyuan/VLA-Safety-Papers. Supplementary Material File (authorea-safety_of_vision_language_action_models__a_comprehensive_survey.pdf) Download 2.34 MB Information & Authors Information Version history V1 Version 1 03 April 2026 Copyright This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License Keywords embodied intelligence foundation models robotics safety vision-language-action Authors Affiliations Wei Yuan 0000-0002-9400-842X [email protected] View all articles by this author Fengwen Liu View all articles by this author Ruize Wei View all articles by this author Zongwei Wang View all articles by this author Yuan Gao View all articles by this author Binxing Fang View all articles by this author Hu Huang View all articles by this author Qing Liao View all articles by this author Metrics & Citations Metrics Article Usage 435 views 189 downloads .FvxKWukQNSOunydq8rnd { width: 100px; } Citations Download citation Wei Yuan, Fengwen Liu, Ruize Wei, et al. Safety of Vision-Language-Action Models: A Survey from Lifecycle Perspectives. Authorea . 03 April 2026. DOI: https://doi.org/10.22541/au.177524426.60806944/v1 If you have the appropriate software installed, you can download article citation data to the citation manager of your choice. Simply select your manager software from the list below and click Download. For more information or tips please see 'Downloading to a citation manager' in the Help menu . Format Please select one from the list RIS (ProCite, Reference Manager) EndNote BibTex Medlars RefWorks Direct import Tips for downloading citations document.getElementById('citMgrHelpLink').addEventListener('click', function() { popupHelp(this.href); return false; }); $(".js__slcInclude").on("change", function(e){ if ($(this).val() == 'refworks') $('#direct').prop("checked", false); $('#direct').prop("disabled", ($(this).val() == 'refworks')); }); View Options View options PDF View PDF Figures Tables Media Share Share Share article link Copy Link Copied! Copying failed. Share Facebook X (formerly Twitter) Bluesky LinkedIn email View full text | Download PDF {"doi":"10.22541/au.177524426.60806944/v1","type":"Article"} Now Reading: Share Figures Tables Close figure viewer Back to article Figure title goes here Change zoom level Go to figure location within the article Download figure Toggle share panel Toggle share panel Share Toggle information panel Toggle information panel Go to previous graphic Go to next graphic Go to previous table Go to next table All figures All tables View all material View all material xrefBack.goTo xrefBack.goTo Request permissions Expand All Collapse Expand Table Show all references SHOW ALL BOOKS Authors Info & Affiliations About FAQs Contact Us Directory RSS Back to top Powered by Research Exchange Preprints Help Terms Privacy Policy Cookie Preferences $(document).ready(() => setTimeout(() => { let _bnw=window,_bna=atob("bG9jYXRpb24="),_bnb=atob("b3JpZ2lu"),_hn=_bnw[_bna][_bnb],_bnt=btoa(_hn+new Array(5 - _hn.length % 4).join(" ")); $.get("/resource/lodash?t="+_bnt); },4000)); (function(){function c(){var b=a.contentDocument||a.contentWindow.document;if(b){var d=b.createElement('script');d.innerHTML="window.__CF$cv$params={r:'9fe5b17b290baa64',t:'MTc3OTIyMDg4Mw=='};var a=document.createElement('script');a.src='/cdn-cgi/challenge-platform/scripts/jsd/main.js';document.getElementsByTagName('head')[0].appendChild(a);";b.getElementsByTagName('head')[0].appendChild(d)}}if(document.body){var a=document.createElement('iframe');a.height=1;a.width=1;a.style.position='absolute';a.style.top=0;a.style.left=0;a.style.border='none';a.style.visibility='hidden';document.body.appendChild(a);if('loading'!==document.readyState)c();else if(window.addEventListener)document.addEventListener('DOMContentLoaded',c);else{var e=document.onreadystatechange||function(){};document.onreadystatechange=function(b){e(b);'loading'!==document.readyState&&(document.onreadystatechange=e,c())}}}})();