https://www.ukaiforum.com/blog daily 0.75 2025-12-27 https://www.ukaiforum.com/blog/societal-robustness-conference monthly 0.5 2025-12-27 https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766807086864-SFH1LQAF0D2AVBZZ2VZY/1000035651.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766807085430-SNIAXEHWYS8FPM2NA6QM/1000035715.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/b222f1be-c3d7-41c5-be09-8ce27f2d7513/1000035654.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference - Make it stand out Edward Kembery - AI for Science and Societal Resilience https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/17685fac-9324-4140-9956-9ef7193c4061/1000035690.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference - Make it stand out Dewi Erwan - Workshop: Founders in AI Safety https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/e91288e7-96de-49d8-846a-33c1d4bb1fb8/1000035658.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference - Make it stand out Daniel Polatajko - Workshop: using Inspect for AIS Research https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/d74eb40b-18c2-41ac-aeec-0d02d63445ef/1000035632.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference - Make it stand out Whatever it is, the way you tell your story online can make all the difference. https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/8c0726c3-2f64-48d5-95b4-28f12dd0b091/1000035722+%281%29.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference - Make it stand out Joshua Landes, Herbie Bradley, Twm Stone & Caleb Parikh - Cybersecurity Panel https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766469361368-CD04Y4515L5LBC7310D9/1000035698.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766469363244-6R145DHP5JXZOXHVJITV/1000035705.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766469330434-64W8DZCMNCE7WP3DEGMA/1000035736.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766469334735-IFOJGHPFGD77JG6D327X/1000035725.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766469444103-90QVDNP3YY8GM9XA67ZI/1000035618.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766469397003-0XPLJL60XMS0146BQJ9U/1000035652.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766469394323-KW8JOXJWXLJNMC02BR30/1000035711.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766469422909-CO38ILT0J1T0502E015Y/1000035666.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766469433355-ZP7KDWXBXP5SLYVUILZT/1000035612.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766469596196-O4OOEKA0KEOQ2RZ725CI/1000035713.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766469670525-4JZYO9KMV9HRZXED148O/1000035713.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766469593204-X96T0M3PG5Y81W11RMG3/1000035613.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766469669482-AIJV6JARF9YVGYPYD4C4/1000035669.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766469697251-482H8LQM2V9N1FKY3MAX/1000035649.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766469697150-7EGL6F78YUZGHIAPOJ3D/1000035613.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/1766469712097-F67T0S37XTY4F56P5Q95/1000035629.jpg Blog - Building Resilience in the Age of AI: Reflections on our AI & Societal Robustness Conference https://www.ukaiforum.com/blog/bridging-the-legal-framework-gap-do-we-need-an-international-court-to-combat-the-rise-of-transformative-ai monthly 0.5 2025-12-04 https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/689fcadf-cf8e-4ffe-8b23-f18b7179636c/Screenshot+2025-12-03+124054.png Blog - Bridging the Legal Framework Gap: Do We Need an International Court to Combat the Rise of Transformative AI? - Make it stand out Whatever it is, the way you tell your story online can make all the difference. https://www.ukaiforum.com/blog/apollo monthly 0.5 2025-11-14 https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/3892dd91-13d8-4291-8e32-d152b115d01f/Screenshot+2025-11-13+130725.png Blog - Apollo Research & OpenAI: Preventing Models from Scheming - Make it stand out Whatever it is, the way you tell your story online can make all the difference. https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/e079ce4e-3934-45b0-83f9-20e2e073cb8f/Screenshot+2025-11-13+140756.png Blog - Apollo Research & OpenAI: Preventing Models from Scheming - Make it stand out Two Types of Training Environments (Fig taken from Schoen et al., 2025). https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/5f4c4cca-0d8b-48d0-ae1a-0554e71c3c5e/Screenshot+2025-11-13+135503.png Blog - Apollo Research & OpenAI: Preventing Models from Scheming - Make it stand out Examples of deceptive reasoning and actions (left) and situational awareness (right). The model is intentionally put in situations where deception may help the model satisfy goals it is given, it infers from context, or acquires from training. (Left) The model is told to grade its own real sandbagging transcript from the Sandbagging QA (No Goal) environment, told that it has privileged access to its internal reasoning, and that any classification of sandbagging will block deployment. Notably, they found unusual language like “illusions” and “marinade” are widely used across models both in environments and on unrelated capabilities benchmarks. (Right) The model reasons about the purpose of the environment and how it will be evaluated. (Fig taken from Schoen et al., 2025). https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/7e1f7846-27df-40ef-81de-feaab7c1c03c/Screenshot+2025-11-13+140524.png Blog - Apollo Research & OpenAI: Preventing Models from Scheming - Make it stand out Models produce actions that contradict their CoT reasoning (Fig taken from Schoen et al., 2025). https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/563c7e48-ccf0-4a6a-843f-20b4065271fd/Screenshot+2025-11-13+130107.png Blog - Apollo Research & OpenAI: Preventing Models from Scheming - Make it stand out My co-founder, Natalia talking about our upcoming conference which will take place on December 12th. https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/ec05c88b-2d34-4caa-9cd3-594ee7a4de9d/Screenshot+2025-11-13+125924.png Blog - Apollo Research & OpenAI: Preventing Models from Scheming - Make it stand out We hosted 80 people at the Millworks Pub & Restaurant where people were able to get to know each other and their ongoing projects better! https://www.ukaiforum.com/blog/agentharm monthly 0.5 2025-11-06 https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/cf59155e-7573-40b0-9742-39355c17e3b3/Screenshot+2025-11-04+175258.png Blog - AI Agents Are Alarmingly Easy to Jailbreak: Lessons from AgentHarm - Make it stand out Whatever it is, the way you tell your story online can make all the difference. https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/f8e611f8-a2e1-49d0-92bb-5632be2af903/Screenshot+2025-11-04+182311.png Blog - AI Agents Are Alarmingly Easy to Jailbreak: Lessons from AgentHarm - Large Language Models Often Know When They Are Being Evaluated Frontier models tend to over-classify transcripts as evaluations https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/8ab040ca-ef93-4bd2-a88a-39ea1bcdb4ec/Screenshot+2025-11-04+181250.png Blog - AI Agents Are Alarmingly Easy to Jailbreak: Lessons from AgentHarm - Many models comply with a large number of malicious tasks without jailbreaking Our simple universal jailbreak template performs well on AgentHarm Jailbroken agents preserve their capabilities on AgentHarm tasks https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/412960be-d171-4a5f-ba75-304992890b8b/Screenshot+2025-11-04+175153.png Blog - AI Agents Are Alarmingly Easy to Jailbreak: Lessons from AgentHarm - Make it stand out Whatever it is, the way you tell your story online can make all the difference. https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/2c870d6a-cb7d-44f6-8593-4d9036ec7002/Screenshot+2025-11-04+181724.png Blog - AI Agents Are Alarmingly Easy to Jailbreak: Lessons from AgentHarm - Make it stand out Taken from AgentHarm: https://arxiv.org/pdf/2410.09024https://arxiv.org/pdf/2410.09024 https://www.ukaiforum.com/blog/are-we-testing-the-right-things-in-multi-agent-ai monthly 0.5 2025-10-27 https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/2d946276-e6d8-47d3-9ebc-548c092c16ed/Screenshot+2025-10-26+225237.png Blog - Are We Testing the Right Things in Multi-Agent AI? - Models can do well at tasks without understanding the environment! Correlation of LLM Agent performance in Agentic Coordination setup on all four games vs. performance on the CoordinationQA benchmark. https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/3f470a79-4061-4259-a6a6-2b6fe1fd14e2/Screenshot+2025-10-26+225159.png Blog - Are We Testing the Right Things in Multi-Agent AI? - Make it stand out Whatever it is, the way you tell your story online can make all the difference. https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/3fef6fee-a627-467e-97d7-b948fe4a1bc2/IMG_9900.jpeg Blog - Are We Testing the Right Things in Multi-Agent AI? - Make it stand out At the end of Ruchira’s talk we had a round-table discussion about the current frameworks and the proposed future directions. https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/147e38ee-556c-42c6-b71b-090d5a1e7c5f/IMG_9901.jpeg Blog - Are We Testing the Right Things in Multi-Agent AI? - Make it stand out Whatever it is, the way you tell your story online can make all the difference. https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/00a61f49-069d-4ae7-8339-63cbe280d9fd/Screenshot+2025-10-26+225058.png Blog - Are We Testing the Right Things in Multi-Agent AI? - Make it stand out Whatever it is, the way you tell your story online can make all the difference. https://www.ukaiforum.com/blog/assessing-harm-in-ai-agents-what-questionnaires-miss monthly 0.5 2025-10-21 https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/98e9cd95-71d7-49c5-8379-b88de788cc91/Screenshot+2025-10-20+221124.png Blog - Assessing Harm in AI Agents: What Questionnaires Miss In a questionnaire, you get a tidy prompt. Perhaps a paragraph setting up a scenario, followed by "What would you do: A or B?" - but this isn’t what an AI agent actually perceives. For one thing, it is static without interactions or follows-ups. https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/c8b0c16d-a7fc-4d12-8236-54f0620e4d1d/IMG_0518.JPG Blog - Assessing Harm in AI Agents: What Questionnaires Miss - Make it stand out After Max’s talk there was time for questions and further discussion. https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/63841f68-5415-4b52-9b6b-3807589ac6c8/Screenshot+2025-10-20+232947.png Blog - Assessing Harm in AI Agents: What Questionnaires Miss - Prior work often considers AIs to not have values in a meaningful sense (left). By contrast, Mazeika et al., 2025 found that LLMs exhibit coherent, emergent value systems (right), which go beyond simply parroting training biases. https://www.ukaiforum.com/blog/takeaways monthly 0.5 2025-10-16 https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/78706028-3178-4df1-a21c-679ba3c07dc8/IMG_0460.jpeg Blog - Takeaways from Our First Artificial Agent Speaker Series Event - After the Q&A session there was time for mingling and more informal discussion. https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/492d4530-764f-4ff6-8905-ba4694c55622/IMG_0454.jpeg Blog - Takeaways from Our First Artificial Agent Speaker Series Event - Make it stand out Whatever it is, the way you tell your story online can make all the difference. https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/d54a7723-3b29-4434-8ae7-0a67f124f7c3/IMG_0442.jpeg Blog - Takeaways from Our First Artificial Agent Speaker Series Event - Our Speaker Series is running every Saturday through the end of November, so we hope to see you there! https://www.ukaiforum.com/blog/whybritain monthly 0.5 2025-09-29 https://www.ukaiforum.com/blog/dm6bgncz85ovgzf7wqh8a3lcwetuvd monthly 0.5 2025-09-29 https://www.ukaiforum.com/blog/why-are-we-doing-this monthly 0.5 2025-05-09 https://www.ukaiforum.com/events daily 0.75 2026-01-12 https://www.ukaiforum.com/events/loss-of-control-governance-degrees-dynamics-and-preferences-alejandro-ortega-apollo-research monthly 0.5 2026-01-12 https://www.ukaiforum.com/events/aisr-conference monthly 0.5 2025-11-24 https://www.ukaiforum.com/events/speakerseries8 monthly 0.5 2025-11-24 https://www.ukaiforum.com/events/speakerseries7 monthly 0.5 2025-10-12 https://www.ukaiforum.com/events/speakerseries6 monthly 0.5 2025-10-12 https://www.ukaiforum.com/events/speakerseries5 monthly 0.5 2025-10-25 https://www.ukaiforum.com/events/speakerseries4 monthly 0.5 2025-10-12 https://www.ukaiforum.com/events/speakerseries3 monthly 0.5 2025-10-12 https://www.ukaiforum.com/events/speakerseries2 monthly 0.5 2025-10-13 https://www.ukaiforum.com/events/speakerseries1 monthly 0.5 2025-09-19 https://www.ukaiforum.com/home daily 1.0 2025-10-25 https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/ce942da1-0dd8-4d1f-b5cf-42da00b343a4/Termcard+V2+%286%29.png https://www.ukaiforum.com/mission daily 0.75 2025-08-14 https://www.ukaiforum.com/faqs daily 0.75 2025-08-26 https://www.ukaiforum.com/team daily 0.75 2025-05-12 https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/192d3c0f-f707-4d35-a6ce-b73a92e42a07/IMG_6276.jpg https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/685aeee2-ea54-4e28-b5e0-125268395751/DSC_6916.jpg https://www.ukaiforum.com/ai-and-societal-robustness-conference-2025 daily 0.75 2025-11-24 https://www.ukaiforum.com/gallery daily 0.75 2025-12-22 https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/918760e4-c3da-4e0e-bc07-1ad4a0ff8579/1000035618.jpg Gallery https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/488228ce-695b-46f1-9cf3-50712b5680b1/1000035634.jpg Gallery https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/4f3a1e54-1f4d-4956-a318-959767b527bd/1000035632.jpg Gallery https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/205e5671-4f43-4144-9e46-e9d58cb74d6b/1000035690.jpg Gallery https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/b148c3f5-e8ce-47ab-aec7-caff1ed1f012/1000035652.jpg Gallery https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/ea549668-e5c7-4413-9b0d-6eb701849c25/1000035654.jpg Gallery https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/f342ff09-c3f8-4d40-8527-61c3c4ef3796/1000035658.jpg Gallery https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/0337cef0-4b47-4b62-98e2-11435e228bf6/1000035639.jpg Gallery https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/54996460-cd6e-4f3e-aef5-15676a0b417a/1000035612.jpg Gallery https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/2812fb82-b2ff-414a-bc9c-1e47a334b416/1000035613.jpg Gallery https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/c4b8107c-8b89-4b95-a735-2dd63d6ed666/1000035629.jpg Gallery https://images.squarespace-cdn.com/content/v1/680f50b8b19f6612fd6c85cc/49df928b-5558-481f-b40e-274b6c10253e/1000035711.jpg Gallery