{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,3]],"date-time":"2026-09-03T06:15:52Z","timestamp":1788416152677,"version":"build-2803163510"},"reference-count":58,"publisher":"Springer Science and Business Media LLC","issue":"8130","license":[{"start":{"date-parts":[[2026,9,2]],"date-time":"2026-09-02T00:00:00Z","timestamp":1788307200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2026,9,2]],"date-time":"2026-09-02T00:00:00Z","timestamp":1788307200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0"}],"content-domain":{"domain":["link.springer.com"],"crossmark-restriction":false},"short-container-title":["Nature"],"published-print":{"date-parts":[[2026,9,3]]},"abstract":"<jats:title>Abstract<\/jats:title>\n                  <jats:p>\n                    Self-driving cars increasingly rely on deep neural networks to achieve human-like driving\n                    <jats:sup>1\u20133<\/jats:sup>\n                    . The opacity of these black-box planners makes it challenging to accurately anticipate when they will fail\n                    <jats:sup>4\u20136<\/jats:sup>\n                    , with potentially catastrophic consequences\n                    <jats:sup>7\u20139<\/jats:sup>\n                    . Although research into interpreting these systems has surged, most of it is confined to simulations or toy setups because of the difficulty of real-world deployment\n                    <jats:sup>10,11<\/jats:sup>\n                    , leaving the practical utility of these techniques unknown. Here, we introduce the Concept-Wrapper Network (CW-Net), a method for faithfully explaining the behaviour of machine-learning-based planners that causally grounds their reasoning in human-interpretable concepts without sacrificing performance. We deploy CW-Net on a real self-driving car and show that the resulting explanations improve the human driver\u2019s mental model of the vehicle, allowing them to better predict its behaviour, particularly in surprising situations. This demonstrates that explainable deep learning integrated into self-driving cars can be both understandable and useful in a realistic deployment setting. We anticipate our method could be applied to other safety-critical systems, such as autonomous drones and robotic surgeons, as well as to other architectures, such as end-to-end learning systems and vision\u2013language\u2013action models. Overall, our study establishes a deployment-validated pathway to interpretability for autonomous agents, which could help make them more transparent and safe.\n                  <\/jats:p>","DOI":"10.1038\/s41586-026-10950-5","type":"journal-article","created":{"date-parts":[[2026,9,2]],"date-time":"2026-09-02T15:03:20Z","timestamp":1788361400000},"page":"114-120","update-policy":"https:\/\/doi.org\/10.1007\/springer_crossmark_policy","source":"Crossref","is-referenced-by-count":0,"title":["Explainable deep learning improves human mental models of self-driving cars"],"prefix":"10.1038","volume":"657","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-5800-2525","authenticated-orcid":false,"given":"Eoin M.","family":"Kenny","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Akshay","family":"Dharmavaram","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Sang Uk","family":"Lee","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Tung","family":"Phan-Minh","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Shreyas","family":"Rajesh","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Yunqing","family":"Hu","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Laura","family":"Major","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-0949-302X","authenticated-orcid":false,"given":"Momchil S.","family":"Tomov","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Julie A.","family":"Shah","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"297","published-online":{"date-parts":[[2026,9,2]]},"reference":[{"key":"10950_CR1","unstructured":"Tomov, M. S. et al. TreeIRL: safe urban driving with tree search and inverse reinforcement learning. In Proc. 2026 IEEE International Conference on Robotics and Automation (ICRA) (in the press)."},{"key":"10950_CR2","doi-asserted-by":"crossref","unstructured":"Phan-Minh, T. et al. Driveirl: drive in real life with inverse reinforcement learning. In Proc. 2023 IEEE International Conference on Robotics and Automation (ICRA), 1544\u20131550 (IEEE, 2023).","DOI":"10.1109\/ICRA48891.2023.10160449"},{"key":"10950_CR3","doi-asserted-by":"crossref","unstructured":"Marcu, A.-M. et al. LingoQA: visual question answering for autonomous driving. In European Conference on Computer Vision (eds. Leonardis, A. et al.) 252\u2013269 (Springer, 2024).","DOI":"10.1007\/978-3-031-72980-5_15"},{"key":"10950_CR4","unstructured":"Nikolaidis, S. & Shah, J. Human-robot teaming using shared mental models. In Proc. ACM\/IEEE Conference on Human\u2013Robot Interaction (IEEE, 2012)."},{"key":"10950_CR5","unstructured":"Major, L. & Shah, J. What to Expect When You\u2019re Expecting Robots: The Future of Human-Robot Collaboration (Hachette, 2020)."},{"key":"10950_CR6","first-page":"610","volume":"34","author":"R Paleja","year":"2021","unstructured":"Paleja, R., Ghuy, M., Ranawaka Arachchige, N., Jensen, R. & Gombolay, M. The utility of explainable AI in ad hoc human-machine teaming. Adv. Neural Inf. Process. Syst. 34, 610\u2013623 (2021).","journal-title":"Adv. Neural Inf. Process. Syst."},{"key":"10950_CR7","unstructured":"Titcomb, J. Uber\u2019s safety policies under fire as US watchdog investigates self-driving car death. The Telegraph. https:\/\/www.telegraph.co.uk\/technology\/2019\/11\/19\/ubers-safety-policies-fire-us-watchdog-investigating-self-driving\/ (10 November 2019)."},{"key":"10950_CR8","unstructured":"Fang, L. Tesla crash footage shows driver with hands off wheel, raising fresh questions about autopilot safety. The Intercept. https:\/\/theintercept.com\/2023\/01\/10\/tesla-crash-footage-autopilot\/ (10 January 2023)."},{"key":"10950_CR9","unstructured":"Templeton, B. Waymo\u2019s double crash with pickup trucks and more examined. Forbes. https:\/\/www.forbes.com\/sites\/bradtempleton\/2024\/03\/05\/waymos-double-crash-with-pickup-trucks-and-more-examined\/ (5 March 2024)."},{"key":"10950_CR10","unstructured":"Kenny, E. M., Tucker, M. & Shah, J. Towards interpretable deep reinforcement learning with human-friendly prototypes. In Proc. Eleventh International Conference on Learning Representations (ICLR, 2023)."},{"key":"10950_CR11","doi-asserted-by":"publisher","first-page":"101603","DOI":"10.1109\/ACCESS.2024.3431437","volume":"12","author":"S Atakishiyev","year":"2024","unstructured":"Atakishiyev, S., Salameh, M., Yao, H. & Goebel, R. Explainable artificial intelligence for autonomous driving: a comprehensive overview and field guide for future research directions. IEEE Access 12, 101603\u2013101625 (2024).","journal-title":"IEEE Access"},{"key":"10950_CR12","doi-asserted-by":"publisher","first-page":"113816","DOI":"10.1016\/j.eswa.2020.113816","volume":"165","author":"C Badue","year":"2021","unstructured":"Badue, C. et al. Self-driving cars: a survey. Expert Syst. Appl. 165, 113816 (2021).","journal-title":"Expert Syst. Appl."},{"key":"10950_CR13","unstructured":"On-Road Automated Driving (ORAD) Committee. Taxonomy and Definitions for Terms Related to Driving Automation Systems for On-Road Motor Vehicles, Tech. Rep. SAE J3016 (SAE International, 2021)."},{"key":"10950_CR14","doi-asserted-by":"publisher","first-page":"103199","DOI":"10.1016\/j.trc.2021.103199","volume":"128","author":"Y Xing","year":"2021","unstructured":"Xing, Y., Lv, C., Cao, D. & Hang, P. Toward human-vehicle collaboration: Review and perspectives on human-centered collaborative automated driving. Transp. Res. C Emerg. Technol. 128, 103199 (2021).","journal-title":"Transp. Res. C Emerg. Technol."},{"key":"10950_CR15","doi-asserted-by":"publisher","first-page":"579","DOI":"10.3390\/make2040031","volume":"2","author":"A Pereira","year":"2020","unstructured":"Pereira, A. & Thomas, C. Challenges of machine learning applied to safety-critical cyber-physical systems. Mach. Learn. Knowl. Extr. 2, 579\u2013602 (2020).","journal-title":"Mach. Learn. Knowl. Extr."},{"key":"10950_CR16","doi-asserted-by":"publisher","first-page":"19342","DOI":"10.1109\/TITS.2024.3474469","volume":"25","author":"A Kuznietsov","year":"2024","unstructured":"Kuznietsov, A., Gyevnar, B., Wang, C., Peters, S. & Albrecht, S. V. Explainable AI for safe and trustworthy autonomous driving: a systematic review. IEEE Trans. Intell. Transp. Syst. 25, 19342\u201319364 (2024).","journal-title":"IEEE Trans. Intell. Transp. Syst."},{"key":"10950_CR17","doi-asserted-by":"crossref","unstructured":"Arfini, S. et al. in Connected and Automated Vehicles: Integrating Engineering and Ethics (eds Fossa, F. & Cheli, F.) 63\u201385 (Springer, 2023).","DOI":"10.1007\/978-3-031-39991-6_4"},{"key":"10950_CR18","doi-asserted-by":"crossref","unstructured":"Atakishiyev, S., Salameh, M. & Goebel, R. Incorporating explanations into human-machine interfaces for trust and situation awareness in autonomous vehicles. In Proc. 2024 IEEE Intelligent Vehicles Symposium (IV), 2948\u20132955 (IEEE, 2024).","DOI":"10.1109\/IV55156.2024.10588812"},{"key":"10950_CR19","doi-asserted-by":"publisher","first-page":"269","DOI":"10.1007\/s12008-014-0227-2","volume":"9","author":"J Koo","year":"2015","unstructured":"Koo, J. et al. Why did my car just do that? Explaining semi-autonomous driving actions to improve driver understanding, trust, and performance. Int. J. Interact. Des. Manuf. 9, 269\u2013275 (2015).","journal-title":"Int. J. Interact. Des. Manuf."},{"key":"10950_CR20","doi-asserted-by":"publisher","first-page":"377","DOI":"10.1504\/IJVD.2016.076740","volume":"70","author":"J Koo","year":"2016","unstructured":"Koo, J., Shin, D., Steinert, M. & Leifer, L. Understanding driver responses to voice alerts of autonomous car operations. Int. J. Veh. Des. 70, 377\u2013392 (2016).","journal-title":"Int. J. Veh. Des."},{"key":"10950_CR21","doi-asserted-by":"crossref","unstructured":"Wiegand, G., Eiband, M., Haubelt, M. & Hussmann, H. \u2018I\u2019d like an explanation for that!\u201d exploring reactions to unexpected autonomous driving. In Proc. 22nd International Conference on Human-Computer Interaction with Mobile Devices and Services, Vol. 36, 1\u201311 (ACM, 2020).","DOI":"10.1145\/3379503.3403554"},{"key":"10950_CR22","doi-asserted-by":"crossref","unstructured":"Wang, C., Weisswange, T. H., Krueger, M. & Wiebel-Herboth, C. B. Human-vehicle cooperation on prediction-level: Enhancing automated driving with human foresight. In Proc. 2021 IEEE Intelligent Vehicles Symposium Workshops (IV Workshops), 25\u201330 (IEEE, 2021).","DOI":"10.1109\/IVWorkshops54471.2021.9669247"},{"key":"10950_CR23","doi-asserted-by":"crossref","unstructured":"Schneider, T., Ghellal, S., Love, S. & Gerlicher, A. R. S. Increasing the user experience in autonomous driving through different feedback modalities. In Proc. 26th International Conference on Intelligent User Interfaces, 7\u201310 (ACM, 2021).","DOI":"10.1145\/3397481.3450687"},{"key":"10950_CR24","doi-asserted-by":"crossref","unstructured":"Schneider, T. et al. Explain yourself! Transparency for positive UX in autonomous driving. In Proc. 2021 CHI Conference on Human Factors in Computing Systems, 161, 1\u201312 (ACM, 2021).","DOI":"10.1145\/3411764.3446647"},{"key":"10950_CR25","doi-asserted-by":"crossref","unstructured":"Omeiza, D., Web, H., Jirotka, M. & Kunze, L. Towards accountability: providing intelligible explanations in autonomous driving. In Proc. 2021 IEEE Intelligent Vehicles Symposium (IV), 231\u2013237 (IEEE, 2021).","DOI":"10.1109\/IV48863.2021.9575917"},{"key":"10950_CR26","doi-asserted-by":"crossref","unstructured":"Zemni, M. et al. OCTET: object-aware counterfactual explanations. In Proc. IEEE\/CVF Conference on Computer Vision and Pattern Recognition, 15062\u201315071 (IEEE, 2023).","DOI":"10.1109\/CVPR52729.2023.01446"},{"key":"10950_CR27","first-page":"1","volume":"7","author":"G Kim","year":"2023","unstructured":"Kim, G., Yeo, D., Jo, T., Rus, D. & Kim, S. What and when to explain? On-road evaluation of explanations in highly automated vehicles. Proc. ACM Interact. Mob. Wearable Ubiquitous Technol. 7, 1\u201326 (2023).","journal-title":"Proc. ACM Interact. Mob. Wearable Ubiquitous Technol."},{"key":"10950_CR28","doi-asserted-by":"crossref","unstructured":"Schneider, T. et al. Don\u2019t fail me! The level 5 autonomous driving information dilemma regarding transparency and user experience. In Proc. 28th International Conference on Intelligent User Interfaces, 540\u2013552 (ACM, 2023).","DOI":"10.1145\/3581641.3584085"},{"key":"10950_CR29","doi-asserted-by":"publisher","unstructured":"Wang, Y. et al. Alpamayo-r1: bridging reasoning and action prediction for generalizable autonomous driving in the long tail. Preprint at https:\/\/doi.org\/10.48550\/arXiv.2511.00088 (2025).","DOI":"10.48550\/arXiv.2511.00088"},{"key":"10950_CR30","unstructured":"Yuksekgonul, M., Wang, M. & Zou, J. Post-hoc concept bottleneck models. In Proc. 11th International Conference on Learning Representations (ICLR, 2023)."},{"key":"10950_CR31","unstructured":"Leake, D. B. (ed.) Case-Based Reasoning: Experiences, Lessons and Future Directions (MIT Press, 1996)."},{"key":"10950_CR32","doi-asserted-by":"crossref","unstructured":"Keane, M. T. & Kenny, E. M. How case-based reasoning explains neural networks: a theoretical analysis of XAI using post-hoc explanation-by-example from a survey of ANN-CBR twin-systems. In Proc. 27th International Conference on Case-Based Reasoning Research and Development ICCBR 2019155\u2013171 (Springer, 2019).","DOI":"10.1007\/978-3-030-29249-2_11"},{"key":"10950_CR33","doi-asserted-by":"publisher","first-page":"109","DOI":"10.1007\/s10462-005-4607-7","volume":"24","author":"F S\u00f8rmo","year":"2005","unstructured":"S\u00f8rmo, F., Cassens, J. & Aamodt, A. Explanation in case-based reasoning\u2013perspectives and goals. Artif. Intell. Rev. 24, 109\u2013143 (2005).","journal-title":"Artif. Intell. Rev."},{"key":"10950_CR34","doi-asserted-by":"crossref","unstructured":"Kenny, E. M. & Keane, M. T. Twin-systems to explain artificial neural networks using case-based reasoning: comparative tests of feature-weighting methods in ann-cbr twins for xai. In Proc. Twenty-Eighth International Joint Conference on Artificial Intelligence (ed. Kraus, S.) (IJCAI, 2019).","DOI":"10.24963\/ijcai.2019\/376"},{"key":"10950_CR35","unstructured":"Schank, R. C. Dynamic Memory: A Theory of Reminding and Learning in Computers and People (Cambridge Univ. Press, 1983)."},{"key":"10950_CR36","first-page":"4768","volume":"30","author":"SM Lundberg","year":"2017","unstructured":"Lundberg, S. M. & Lee, S.-I. A unified approach to interpreting model predictions. Adv. Neural Inf. Process Syst. 30, 4768\u20134777 (2017).","journal-title":"Adv. Neural Inf. Process Syst."},{"key":"10950_CR37","doi-asserted-by":"crossref","unstructured":"Ribeiro, M. T., Singh, S. & Guestrin, C. \u201cWhy should I trust you?\u201d explaining the predictions of any classifier. In Proc. 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 1135\u20131144 (ACM, 2016).","DOI":"10.1145\/2939672.2939778"},{"key":"10950_CR38","doi-asserted-by":"publisher","first-page":"206","DOI":"10.1038\/s42256-019-0048-x","volume":"1","author":"C Rudin","year":"2019","unstructured":"Rudin, C. Stop explaining black box machine learning models for high stakes decisions and use interpretable models instead. Nat. Mach. Intell. 1, 206\u2013215 (2019).","journal-title":"Nat. Mach. Intell."},{"key":"10950_CR39","doi-asserted-by":"crossref","unstructured":"Karnchanachari, N. et al. Towards learning-based planning: The nuplan benchmark for real-world autonomous driving. In Proc. 2024 IEEE International Conference on Robotics and Automation (ICRA), 629\u2013636 (ICRA, 2024).","DOI":"10.1109\/ICRA57147.2024.10610077"},{"key":"10950_CR40","doi-asserted-by":"crossref","unstructured":"Heim, M., Su\u00e1rez-Ruiz, F., Bhuiyan, I., Brito, B. & Tomov, M. S. Lab2car: A versatile wrapper for deploying experimental planners in complex real-world environments. In Proc. 2025 IEEE International Conference on Robotics and Automation (ICRA), 14828\u201314834 (IEEE, 2025).","DOI":"10.1109\/ICRA55743.2025.11128008"},{"key":"10950_CR41","first-page":"44","volume":"40","author":"D Gunning","year":"2019","unstructured":"Gunning, D. & Aha, D. Darpa\u2019s explainable artificial intelligence (XAI) program. AI Mag. 40, 44\u201358 (2019).","journal-title":"AI Mag."},{"key":"10950_CR42","unstructured":"Foster, M. I. & Keane, M. T. Surprise! You\u2019ve got some explaining to do. In Proc. Annual Meeting of the Cognitive Science Society, Vol. 35, 2321\u20132326 (Cognitive Science Society, 2013)."},{"key":"10950_CR43","doi-asserted-by":"publisher","first-page":"1096257","DOI":"10.3389\/fcomp.2023.1096257","volume":"5","author":"RR Hoffman","year":"2023","unstructured":"Hoffman, R. R., Mueller, S. T., Klein, G. & Litman, J. Measures for explainable AI: explanation goodness, user satisfaction, mental models, curiosity, trust, and human-AI performance. Front. Comput. Sci. 5, 1096257 (2023).","journal-title":"Front. Comput. Sci."},{"key":"10950_CR44","doi-asserted-by":"publisher","first-page":"65","DOI":"10.1518\/001872095779049499","volume":"37","author":"MR Endsley","year":"1995","unstructured":"Endsley, M. R. Measurement of situation awareness in dynamic systems. Hum. Factors 37, 65\u201384 (1995).","journal-title":"Hum. Factors"},{"key":"10950_CR45","unstructured":"Federal Aviation Administration. Flight Test Guide for Certification of Part 23 Airplanes. Advisory Circular AC No. 23-8C, US Department of Transportation Initiated By: ACE-100 (2011)."},{"key":"10950_CR46","doi-asserted-by":"publisher","first-page":"1772","DOI":"10.1080\/10447318.2022.2081282","volume":"38","author":"L Sanneman","year":"2022","unstructured":"Sanneman, L. & Shah, J. A. The situation awareness framework for explainable AI (SAFE-AI) and human factors considerations for XAI systems. Int. J. Hum. Comput. Interact. 38, 1772\u20131788 (2022).","journal-title":"Int. J. Hum. Comput. Interact."},{"key":"10950_CR47","doi-asserted-by":"publisher","first-page":"939","DOI":"10.1016\/j.ergon.2005.04.002","volume":"35","author":"R Ma","year":"2005","unstructured":"Ma, R. & Kaber, D. B. Situation awareness and workload in driving while using adaptive cruise control and a cell phone. Int. J. Ind. Ergon. 35, 939\u2013953 (2005).","journal-title":"Int. J. Ind. Ergon."},{"key":"10950_CR48","doi-asserted-by":"publisher","first-page":"450","DOI":"10.1109\/TSMCA.2005.850589","volume":"35","author":"JC Scholtz","year":"2005","unstructured":"Scholtz, J. C., Antonishek, B. & Young, J. D. Implementation of a situation awareness assessment tool for evaluation of human-robot interfaces. IEEE Trans. Syst. Man Cybern. A Syst. Hum. 35, 450\u2013459 (2005).","journal-title":"IEEE Trans. Syst. Man Cybern. A Syst. Hum."},{"key":"10950_CR49","doi-asserted-by":"publisher","first-page":"103459","DOI":"10.1016\/j.artint.2021.103459","volume":"294","author":"EM Kenny","year":"2021","unstructured":"Kenny, E. M., Ford, C., Quinn, M. & Keane, M. T. Explaining black-box classifiers using post-hoc explanations-by-example: The effect of explanations and error-rates in XAI user studies. Artif. Intell. 294, 103459 (2021).","journal-title":"Artif. Intell."},{"key":"10950_CR50","doi-asserted-by":"publisher","first-page":"684","DOI":"10.1037\/0033-295X.102.4.684","volume":"102","author":"G Gigerenzer","year":"1995","unstructured":"Gigerenzer, G. & Hoffrage, U. How to improve Bayesian reasoning without instruction: frequency formats. Psychol. Rev. 102, 684\u2013704 (1995).","journal-title":"Psychol. Rev."},{"key":"10950_CR51","doi-asserted-by":"crossref","unstructured":"Zhou, Z., Ye, L., Wang, J., Wu, K. & Lu, K. HiVT: hierarchical vector transformer for multi-agent motion prediction. In Proc. IEEE\/CVF Conference on Computer Vision and Pattern Recognition, 8823\u20138833 (IEEE, 2022).","DOI":"10.1109\/CVPR52688.2022.00862"},{"key":"10950_CR52","doi-asserted-by":"crossref","unstructured":"Lin, T.-Y., Goyal, P., Girshick, R., He, K. & Doll\u00e1r, P. Focal loss for dense object detection. In Proc. IEEE International Conference on Computer Vision, 2980\u20132988 (IEEE, 2017).","DOI":"10.1109\/ICCV.2017.324"},{"key":"10950_CR53","unstructured":"Schrodi, S., Schur, J., Argus, M. & Brox, T. Selective concept bottleneck models without predefined concepts. Preprint at OpenReview.net https:\/\/openreview.net\/forum?id=PMO30TLI4l (2025)."},{"key":"10950_CR54","unstructured":"Templeton, A. et al. Scaling monosemanticity: extracting interpretable features from Claude 3 Sonnet. Transformer Circuits Thread. https:\/\/transformer-circuits.pub\/2024\/scaling-monosemanticity\/index.html (2024)."},{"key":"10950_CR55","doi-asserted-by":"publisher","first-page":"30046","DOI":"10.1073\/pnas.1907367117","volume":"117","author":"CD Manning","year":"2020","unstructured":"Manning, C. D., Clark, K., Hewitt, J., Khandelwal, U. & Levy, O. Emergent linguistic structure in artificial neural networks trained by self-supervision. Proc. Natl Acad. Sci. USA 117, 30046\u201330054 (2020).","journal-title":"Proc. Natl Acad. Sci. USA"},{"key":"10950_CR56","unstructured":"Bereska, L. & Gavves, S. Mechanistic interpretability for AI safety \u2013 a review. Transactions on Machine Learning Research. https:\/\/leonardbereska.github.io\/blog\/2024\/mechinterpreview\/ (2024)."},{"key":"10950_CR57","doi-asserted-by":"publisher","unstructured":"Marks, S. et al. Auditing language models for hidden objectives. Preprint at https:\/\/doi.org\/10.48550\/arXiv.2503.10965 (2025).","DOI":"10.48550\/arXiv.2503.10965"},{"key":"10950_CR58","doi-asserted-by":"crossref","unstructured":"Endsley, M. R. Direct measurement of situation awareness: Validity and use of SAGAT. In Situational Awareness, 129\u2013156 (Routledge, 2017).","DOI":"10.4324\/9781315087924-9"}],"container-title":["Nature"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/www.nature.com\/articles\/s41586-026-10950-5.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41586-026-10950-5","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/www.nature.com\/articles\/s41586-026-10950-5.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,9,3]],"date-time":"2026-09-03T05:43:08Z","timestamp":1788414188000},"score":1,"resource":{"primary":{"URL":"https:\/\/www.nature.com\/articles\/s41586-026-10950-5"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,9,2]]},"references-count":58,"journal-issue":{"issue":"8130","published-print":{"date-parts":[[2026,9,3]]}},"alternative-id":["10950"],"URL":"https:\/\/doi.org\/10.1038\/s41586-026-10950-5","relation":{},"ISSN":["0028-0836","1476-4687"],"issn-type":[{"value":"0028-0836","type":"print"},{"value":"1476-4687","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,9,2]]},"assertion":[{"value":"24 November 2024","order":1,"name":"received","label":"Received","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"24 July 2026","order":2,"name":"accepted","label":"Accepted","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"2 September 2026","order":3,"name":"first_online","label":"First Online","group":{"name":"ArticleHistory","label":"Article History"}},{"value":"Several authors were employed by Motional during the course of this research (A.D., S.U.L., T.P.-M., S.R., Y.H., L.M. and M.S.T.). Moreover, E.M.K. received compensation from Motional as a research intern during the data collection phase of the project. Certain authors may hold equity interests in Motional, the value of which could potentially be affected by the publication of this research. Motional has pending patent protection regarding certain aspects involving the ML architectures and AV planning systems described in this manuscript (see WIPO Patent App. Pub. No. WO 2026\/085086 (published 23 April 2026)). Authors listed as inventors on this application include E.M.K., M.S.T., A.D., S.U.L., T.P.-M., S.R., Y.H. and J.A.S.","order":1,"name":"Ethics","label":"Competing interests","group":{"name":"EthicsHeading","label":"Ethics"}}]}}