Database

No summary available yet.

No summary available yet.

TruthfulAI is a non-profit AI safety research organization based in Berkeley that studies situational awareness, deception, and hidden reasoning in large language models.

No summary available yet.

No summary available yet.

Luca De Leo is a Co-founding Director of BAISH (Buenos Aires AI Safety Hub) and an AI safety community builder based in Buenos Aires, Argentina. He studied Computer Science at the University of Buenos Aires (UBA), leaving the master's program to focus full-time on AI safety work after receiving grants from ACX+ and the Long-Term Future Fund to develop relevant research skills. He worked in operations at Nonlinear and facilitated AGI Safety Fundamentals courses, growing the AI Safety community at UBA to over 130 active members. He has achieved multiple competitive placements in Apart Research sprints and participated in AI safety hackathons organized at EAGx LatAm. His current work includes co-directing BAISH, serving as a facilitator for BlueDot Impact's AGI Strategy course, and part-time operations for the AI Species YouTube channel.

No summary available yet.

Sienka Dounia is a technical AI safety researcher originally from Chad, currently based in London after relocating to work on Eliciting Latent Knowledge (ELK) research with Jake Mendel at Apollo Research. She is a 2023 fellow of the ILINA Program (Initiative for Longtermism in Africa) and a 2024 fellow of the AI Futures Fellowship in Mexico, where she worked on developmental interpretability and singular learning theory. She is an AI Safety Content Associate and Knowledge & Systems Lead at Successif, a platform supporting professionals transitioning into AI risk reduction roles. She is also a researcher at the African Hub on AI Safety, Peace and Security at the University of Cape Town, and a co-author of the 2025 paper "Assessing the Case for Africa-Centric AI Safety Evaluations," which develops a risk taxonomy for frontier AI deployments in African contexts. Her research interests include model evaluation, model interpretability, AI deception, and the technical governance of advanced AI systems.

Geodesic Research is a technical AI safety organization based in Cambridge, UK, focused on implementing and measuring pre- and post-training methods to improve model safety and alignment.

No summary available yet.

Policy researcher at the Existential Risk Observatory with a background in regulatory policy in government and a tertiary qualification in economics, preparing submissions and policy proposals on AI existential risk.

Program Director for Talk to the City at the AI Objectives Institute, focused on building technology for knowledge curation and large-scale qualitative data. Before joining AOI she managed a DARPA research portfolio on how machines can help humans understand complex geopolitical systems and holds a PhD in computational modeling from the University of Florida and a B.S. in Symbolic Systems from Stanford.

Co-founder and founding advisor of Sage who led software development for the early Quantified Intuitions calibration and pastcasting tools and serves as a director of Sage Future Inc on corporate filings.

No summary available yet.

No summary available yet.

No summary available yet.

Senior Campaigner at the Existential Risk Observatory, founder of Conjointly and the Campaign for AI Safety, with a background in marketing and management consulting and a focus on AI safety advocacy.

A research organization focused on reducing risks of astronomical suffering (s-risks) from advanced AI, with emphasis on conflict prevention and cooperation between transformative AI systems.

No summary available yet.

Andy Graham is a Chartered Mechanical Engineer and Fellow of the Institution of Mechanical Engineers who co-founded Amodo Design, a Sheffield-based multi-disciplinary engineering consultancy, after earning a first-class Mechanical Engineering degree from the University of Sheffield.

Kathleen Fisher is Chief Executive Officer of the Advanced Research and Invention Agency (ARIA). She previously led DARPA’s Information Innovation Office, overseeing more than 50 programmes, and founded a centre at RAND applying AI and formal methods to cybersecurity. She holds a PhD in computer science from Stanford and is a Fellow of AAAS and ACM.

No summary available yet.

Marcus Williams is an AI safety researcher currently working on deception and scheming monitoring at OpenAI. He completed his Master's degree in engineering physics and machine learning at Lund University in 2023. After graduating, he worked at AI Safety Hub Oxford on a theoretical reinforcement learning project, co-authoring "On the Expressivity of Objective-Specification Formalisms in Reinforcement Learning," which was accepted at ICLR 2024. He received a Long-Term Future Fund grant for a six-month independent project on Multi-Objective Reinforcement Learning from AI Feedback (MORLAIF), which produced an arXiv paper demonstrating that decomposing preference modeling into multiple principles outperforms standard RLAIF baselines. In the MATS Summer 2024 cohort under mentor Micah Carroll, he researched annotator vulnerabilities and LLM influence on human preferences, resulting in the paper "Targeted Manipulation and Deception Emerge in LLMs Trained on User Feedback" (accepted at NeurIPS workshops). He is also a co-author on "Stress Testing Deliberative Alignment for Anti-Scheming Training" alongside researchers from OpenAI and Apollo Research.

Researcher at Anthropic whose prior roles include researcher at Google and PhD student at the Technical University of Munich; his expertise spans AI safety and alignment as well as graph neural networks and machine learning on molecules.

Vael Gates is a researcher and field-builder focused on improving the safety of AI systems to reduce potential large-scale risks from advanced AI. They completed a PhD in Computational Cognitive Science at UC Berkeley in Tom Griffiths's lab, working on computational models of social cognition, and subsequently held a postdoctoral fellowship at Stanford University jointly with HAI (Human-Centered AI Institute) and CISAC (Center for International Security and Cooperation). During their Stanford postdoc, they conducted structured interviews with nearly 100 AI researchers about their perceptions of risks from current and future AI systems, publishing results as the "Risks from Advanced AI" (2022) and "Risks from Highly-Capable AI" (2023) series on the EA Forum and LessWrong. They previously founded Arkose, an AI safety field-building nonprofit providing informational resources and support calls to machine learning researchers and engineers interested in entering the field (closed June 2025). They subsequently served as Head of Content at FAR.AI, leading content programming for the FAR.Futures division including global conferences, workshops, and community engagement. As of early 2026, Vael is the Founder and Executive Director of Humans in Control, a bipartisan grassroots organization focused on protecting communities from risks of unchecked AI.

Executive Director of Saving Humanity from Homo Sapiens and a mathematics graduate whose work and interests shifted from academia toward existential risk and philanthropy.

Founder of Aether, an independent research lab focused on foundation model agent safety, and PhD student in computer science at the University of Toronto (currently on leave), working on AI safety and LLM agents.

Liron Shapira is an entrepreneur and founder/CEO of Relationship Hero, a dating and relationship coaching platform, and host of DoomDebates.com.

Wes Ezzeddine is a Dubai-based technology leader and AI safety specialist with over 15 years of experience leading engineering teams across fintech, insurtech, mobility, and edtech. He has served as Director of Engineering at Mamo Pay, co-founded AI Safety UAE, and works on AI alignment, governance, and resilient systems while facilitating AI safety courses and events.

No summary available yet.

Charlie Griffin is a doctoral student in Computer Science at the University of Oxford, affiliated with Green Templeton College and supervised by Alessandro Abate and Marta Kwiatkowska. His research focuses on AI control — developing formal frameworks and safety evaluations for containing misaligned AI systems. He is affiliated with the UK AI Security Institute (AISI) and FAR.AI, where he has contributed research on AI-control games and untrusted monitoring. His notable publications include "Games for AI Control: Models of Safety Evaluations of AI Deployment Protocols" (with Buck Shlegeris and others) and "When can we trust untrusted monitoring? A safety case sketch across collusion strategies" (with researchers from Google DeepMind and AISI). He helped organize the first round of LASR Labs (London AI Safety Research) and continues as an advisor, and has received funding from the Long-Term Future Fund for alignment work including skilling up, assisting academics, personal research, and community building.

A leading U.S. law school that conducts research on AI governance, policy, and safety through its PULSE program and Institute for Technology, Law & Policy.

Professor Rafael A. Calvo is Professor at the Dyson School of Design Engineering at Imperial College London and co‑lead of the Imperial College spoke of the Leverhulme Centre for the Future of Intelligence. His research focuses on the design of digital technologies that support psychological wellbeing, mental health and education, and on the ethical challenges raised by new technologies. He is co‑author of the book Positive Computing and has served as co‑editor of the IEEE Transactions on Technology and Society and on the editorial boards of several leading journals in affective computing and learning technologies.

Joanna Wiaterek is co-founder of the Centre for AI Security and Access with expertise in international AI governance coordination. She has worked on AI benefit-sharing, global AI safety, and foreign aid and is focused on fostering meaningful and inclusive AI futures.

Viktor Rehnberg is a Swedish AI safety researcher based in Gothenburg, Sweden. He holds an M.Sc. in Engineering Physics from Chalmers University of Technology, where he also worked as a Research Engineer at Chalmers e-Commons supporting ML/AI infrastructure. He participated in the SERI MATS (ML Alignment Theory Scholars) Winter 2022 program, conducting research on identifying key steps in reducing risks from learned optimization, including mesa-optimization and inner alignment problems. He has collaborated with Erik Jenner and Oliver Daniels-Koch on empirical mechanistic anomaly detection (MAD) research, co-authoring a LessWrong post on concrete empirical research projects in that area under supervision of John Wentworth and Erik Jenner. He also participated in AI Safety Camp Edition 5, where his team investigated neural network modularity loss functions to improve interpretability. He is an organizer of EA Gothenburg and is motivated by effective altruism, longtermism, and preventing existential risk.

Jason Gross is a computer scientist and entrepreneur, co‑founder of Theorem (Theorem Labs), an AI and programming languages research lab focused on program verification; he holds a PhD in Electrical Engineering and Computer Science from MIT, where his research improved proof assistants and contributed verified cryptography now securing large volumes of HTTPS traffic.

Ajeya Cotra works at METR on threat modeling and risk assessment for loss-of-control risks from advanced AI. She previously led the technical AI safety program at Open Philanthropy (now Coefficient Giving), where she developed the influential “biological anchors” framework for forecasting when transformative AI might arrive.

Aaquib Syed is a CS and Mathematics undergraduate student at the University of Maryland, College Park, where he is a Banneker-Key Scholar. He was a fellow in the MATS 5.0 program under Neel Nanda's supervision, conducting mechanistic interpretability research on how refusal is implemented in large language models. His most notable work, "Refusal in Language Models Is Mediated by a Single Direction" (NeurIPS 2024), co-authored with Andy Arditi and others, showed that refusal behavior across 13 open-source chat models is controlled by a single direction in the residual stream. He also co-authored "Attribution Patching Outperforms Automated Circuit Discovery" (NeurIPS 2024 ATTRIB workshop) and work on mechanistic unlearning and model pruning. He is currently a Student Researcher at Google DeepMind on the Frontier Safety team, where he focuses on evaluating and forecasting dangerous AI capabilities.

No summary available yet.

No summary available yet.

No summary available yet.

Finan Adamson is a rationalist, effective altruist, and activist focused on reducing global catastrophic risks. He holds a BS in Evolutionary Biology from The Evergreen State College and has previously worked as a Lab Tech at the University of Alaska Fairbanks researching methane-digesting bacteria. He served as a communication coordinator at ALLFED (Alliance to Feed the Earth in Disasters), where he became their first full-time employee after volunteering under a Centre for Effective Altruism grant, and also ran events for Seattle Effective Altruists. He worked with the Machine Intelligence Research Institute (MIRI) and served as Events Coordinator and Community Manager at the Foresight Institute. His biosecurity work includes co-authoring the EA Forum post "Bioinfohazards" with Megan Crawford and Jeffrey Ladish, authoring a Nuclear Preparedness Guide, co-organizing Catalyst (a 100-person biosecurity summit), and completing an ORISE fellowship at the Oak Ridge Institute for Science and Education focused on holistic evaluation of field water surveillance and waterborne pathogen detection. He received a Long-Term Future Fund grant to upskill in biosecurity and explore possible career paths in the field.

No summary available yet.

No summary available yet.

No summary available yet.

Robert Miles (also known as Rob Miles) is a British science communicator and educator specializing in AI safety and alignment. He studied Computer Science at the University of Nottingham and dropped out of his PhD program around 2011 to focus on AI safety communication full time. He began creating content for the popular Computerphile YouTube channel around 2015 before launching his own channel, Robert Miles AI Safety, which has accumulated over 169,000 subscribers and more than 7.5 million views covering topics such as the orthogonality thesis, instrumental convergence, and inner misalignment. He is the founder of AISafety.info (also known as Stampy), a community-written interactive FAQ about AI existential risk, and has run the Distillation Fellowship, a paid program funding writers to distill AI safety research into accessible content for the site. He has co-produced the Alignment Newsletter Podcast with Rohin Shah and has received funding from the Long-Term Future Fund to support his educational work. He collaborates with organizations including MIRI and the Future of Humanity Institute to help communicate their research to broader audiences.

No summary available yet.

German computer scientist and leading machine learning researcher, Scientific Director of the ELLIS Institute Tübingen and Director of the Empirical Inference Department at the Max Planck Institute for Intelligent Systems, whose work focuses on machine learning and causal inference with applications ranging from astronomy and computational photography to robotics.

A scientific diplomacy organization working to improve global catastrophic risk governance in Spanish-speaking countries, with focus areas spanning AI regulation, pandemic biosecurity, food security, and risk management systems.