Database

Founder of Frame Fellowship and Explainable (Explainable Media), working on AI safety media infrastructure and field‑building from San Francisco, California.

Oliver Klingefjord is a researcher at the Meaning Alignment Institute, which he co-founded in 2023 with early support from OpenAI. His work focuses on aligning AI and institutions with what people actually value—a problem the team calls full-stack alignment—and he previously worked on AI-assisted democracy at the AI Objectives Institute and co-founded the startup Potential.

4-month stipend for a career transition period to explore roles in AI safety communications

No summary available yet.

Evan Knight is the founder of Thoughtful Content and serves as an advisor to Good Impressions, drawing on entrepreneurial and operational experience from co-founding and serving as COO of the marketing analytics firm RightMetric and advising companies such as Casting Workbook and Kinzoo Technologies. He has a background in digital marketing, e-commerce, and operations and holds a commerce degree from the University of Calgary.

Associate Professor in Computer Science at the University of Birmingham, where he leads the AI Verification Group, and head of Zeroth Research; his work bridges formal methods and AI to develop verification techniques for the safety assurance of software, cyber-physical, and deep learning systems.

Empirical AI Safety Research

No summary available yet.

AI policy researcher and AI risk analyst at ORCG, specializing in international security and technology governance, with prior research roles at OpenAI and Georgetown’s Center for Security and Emerging Technology.

Short Documentary and Music Video

Support for alignment theory agenda evaluation

3 month salary for AI safety work on deconfusion and technical alignment.

1-year stipend (and travel and equipment expenses) for support for work on 2 AI safety projects: 1) Penalising neural networks for learning polysemantic neurons; and 2) Crowdsourcing from volunteers for alignment research.

Co-Founder of Cambridge Biosecurity Hub

No summary available yet.

Import AI is a weekly newsletter by Jack Clark (co-founder of Anthropic) covering cutting-edge AI research and its societal implications, read by over 116,000 subscribers.

Previously did GovAI's summer research fellowship, ran the Supervised Program for Alignment Research, and worked in communications.

No summary available yet.

Associate Professor of Computer Science and Systems & Information Engineering at the University of Virginia and member of the Link Lab whose research focuses on runtime safety and trust for embodied AI systems, developing architectural and algorithmic foundations for safe, adaptive human–AI collaboration in domains such as autonomous driving, robotics, healthcare and smart infrastructure.

No summary available yet.

Master's degree in Cognitive Science. Running the Effectively Handling Disagreements workshop

No summary available yet.

No summary available yet.

Volunteer researcher at the Existential Risk Observatory and postdoctoral mathematician at Utrecht University, interested in AI as well as science and education policy.

No summary available yet.

AI Safety research ML engineer

No summary available yet.

Kairos is a US nonprofit that accelerates talent into AI safety and policy by running university group support programs and research mentorship fellowships.

No summary available yet.

Dr. Merve Ayyüce Kızrak is an AI specialist with over 15 years of experience in artificial intelligence, specialising in computer vision, AI governance, and AI safety. She holds a PhD in Electronics and Communication Engineering from Yıldız Technical University and a master’s in Economics and FinTech from Bahçeşehir University, and works as an AI Specialist in the Presidency of Türkiye’s Digital Transformation Office, where she leads data‑governance projects and contributes to the national AI strategy.

Zohreh Shams is a research manager at MATS London and an affiliated lecturer at the University of Cambridge whose research focuses on interpretability and its applications to knowledge discovery. The MATS London profile notes that she previously served as CSO at Leap Laboratories, a spin-out of MATS, where she led research.

Boston Nyer is a co‑founder and fractional COO of Juniper Ventures and an operator focused on AI and global health, with prior leadership roles at FAR.AI, GiveWell, Equalize Health, Momentum and clean‑cookstove company BURN Manufacturing.

Development Director at Sinergia Animal, where I get to use my international business expertise with a deep commitment and love to animals.

David Williams-King is an AI safety researcher and communicator based in Montreal, Canada. He holds a PhD in Computer Science (Systems & Security) from Columbia University (2014-2020), where his research focused on binary security and code randomization. After graduating, he co-founded and served as CTO of Elpha Secure, a cybersecurity insurance startup, leading a team of 15-20 engineers. He transitioned into AI safety work and currently conducts research at Mila (Quebec AI Institute) as part of Yoshua Bengio's Safe AI for Humanity initiative, and serves as a Research Manager for Technical AI Safety at ERA Cambridge (ERA Fellowship). He also works with LawZero, an AI safety organization. His research examines LLM safety, jailbreaks, and guardrails, drawing lessons from cybersecurity history, and he has co-authored papers including "Can Safety Fine-Tuning Be More Principled? Lessons Learned from Cybersecurity" and "Superintelligent Agents Pose Catastrophic Risks: Can Scientist AI Offer a Safer Path?" with Yoshua Bengio and others. He also runs a YouTube channel with over 30,000 subscribers focused on AI safety and how AI will impact society, and received an LTFF grant to support this work.

No summary available yet.

No summary available yet.

A Coherence based Emergent Protocol

No summary available yet.

No summary available yet.

No summary available yet.

Doom Debates is a podcast and debate show hosted by Liron Shapira focused on high-stakes debates about AI existential risk. Its mission is to raise mainstream awareness of potential extinction from AGI and build social infrastructure for high-quality public discourse on the topic.

No summary available yet.

No summary available yet.

Roman Leventov is an independent researcher, blogger, and philosopher focused on intelligence and agency, with particular emphasis on Active Inference, AI alignment, ethics, and collective decision-making systems. He holds a technical background in software engineering from Bauman Moscow State Technical University in Moscow and has contributed to high-performance open-source systems including Apache Druid. He has worked as a software engineer at companies including Metamarkets and Pocketdata.AI, and runs the "Engineering Ideas" Substack covering systems engineering, AI, and philosophy of agency. His AI safety research examines goal misgeneralisation through the lens of Active Inference and the development of collective sense-making architectures. He is a founding member of the Gaia Consortium, a project aimed at building a global decentralized system for collective decision-making and civilisational intelligence. He received a grant from the Long-Term Future Fund to support six months of independent research on goal misgeneralisation from an Active Inference perspective and collective decision-making systems.

No summary available yet.

Part-time Program Manager at the Berkeley Existential Risk Initiative since July 2025, overseeing new program development. He is a consultant, program manager, and community builder with a background in workforce and economic development, and a trained engineer who has worked in the aerospace and manufacturing industry.

No summary available yet.

Practicing Embodied Protocols that work with Live Interfaces