Database

Max Kaufmann is a PhD student at the University of Toronto supervised by Roger Grosse, and a Visiting Student Researcher at Google DeepMind where he works on chain-of-thought monitoring for AI safety. He completed his undergraduate degree in computer science at the University of Cambridge. Before his PhD, he was part of the founding team at the UK AI Safety Institute, where he helped scale the organisation and established their LLM agents team. He has also worked as a research intern at UC Berkeley's Human-Compatible AI group, collaborating with Owain Evans on LLM generalization research, including co-authoring papers on the Reversal Curse and situational awareness in LLMs. His research spans adversarial robustness, LLM generalization, LLM evaluation, training data attribution, and chain-of-thought monitoring. He has received funding from the Long-Term Future Fund to support his early-career alignment research.

No summary available yet.

General support of flexHEG prototyping

No summary available yet.

No summary available yet.

PhD student in the Algorithmic Alignment Group at MIT whose work uses interpretability and causal methods to explain model and human behaviours, including how people interact with AI-generated content and audits of black-box social media algorithms.

Johannes Heidecke is the Head of Safety Systems at OpenAI, where he leads work on model safety behavior and alignment. He completed a Master's degree in Artificial Intelligence in Barcelona and participated in the MATS Summer 2022 Cohort under the mentorship of Evan Hubinger. Early in his career he organized the second AI Safety Camp in Prague, a retreat for nearly 30 aspiring AI alignment researchers, and received funding to support this field-building work. At OpenAI he has co-authored influential safety research including "Deliberative Alignment: Reasoning Enables Safer Language Models" (2024) and "Improving Model Safety Behavior with Rule-Based Rewards," both of which have shaped how OpenAI's o-series models handle safety-critical decisions. He has spoken at the 2025 Singapore Conference on AI and has been quoted on OpenAI's preparedness framework and the risks posed by advanced reasoning models.

Justin Rockefeller is a philanthropist and impact investing leader who co-founded and chairs The ImPact, a global community of families committed to aligning their assets with their values. He serves on the board and the investment and governance committees of the Rockefeller Brothers Fund and is a board member of Longview Philanthropy; previously he worked at Addepar and in venture capital.

No summary available yet.

No summary available yet.

Shelby Summerfield is an operations and leadership professional with a diverse background in political campaigns, social impact, and startup operations; she serves as Head of Operations at Halcyon Futures (2023–present), is Chief of Staff and Senior Operations Manager at Verbal+Visual, and previously was Chief Operating Officer at Humanity Forward and held senior scheduling and campaign roles for Andrew Yang’s presidential and New York City mayoral campaigns.

Dr. Roxy Mathew Koll is a climate scientist at the Indian Institute of Tropical Meteorology whose research focuses on Indian Ocean warming, monsoon dynamics, and extreme weather in the Indo-Pacific, and who has contributed to IPCC reports and received major honors including India’s 2024 Rashtriya Vigyan Puraskar and the AGU Devendra Lal Medal.

No summary available yet.

No summary available yet.

Talent Partner at the UK AI Security Institute, bringing recruiting experience from previous roles at Chapter 2, Sainsbury’s, John Lewis Partnership and The Crown Estate.

No summary available yet.

Fred Heiding is a postdoctoral research fellow at the Harvard Kennedy School’s Belfer Center for Science and International Affairs. His work focuses on computer security at the intersection of technical capabilities, business implications, and policy remediations. Fred is a member of the World Economic Forum's Cybercrime Center and is on the organizing committee of the Technology and National Security Conference.

Accepted ICML 2026 workshop paper on cross-constitution drift in LLMs; seeking $2,050 travel support to present in Seoul and gather research feedback.

Seeking funding to develop and evaluate a new benchmark for systematically assesing safety of LLM-based agents

Harrison Gietz is an AI Program Associate at Longview Philanthropy, where he conducts grant investigations focused on artificial intelligence. He previously served as Co-Director of the Cambridge AI Safety Hub (CAISH) and Program Director of the ERA Fellowship, collectively supporting over 40 research projects in technical AI safety and AI governance. His background includes published research in machine learning robustness and experimental physics, and he has participated in research programs at MATS and the Center for AI Safety. He holds a Bachelor of Science in Mathematics (summa cum laude) from Louisiana State University and also studied computer science at Georgia Tech. Earlier in his career, he received a Long-Term Future Fund grant for full-time AI safety technical and governance research. He is based in Cambridge, UK, and writes on AI safety topics at his personal Substack.

Joshua Reiners is an AI safety researcher focused on mechanistic interpretability. He received a grant from the Long-Term Future Fund to spend four months investigating the most interpretable directions in GPT-2-small's early residual stream, a project aimed at improving our understanding of how language models represent and process information in their early layers. His work sits within the broader mechanistic interpretability research agenda, which seeks to reverse-engineer neural network computations into human-understandable algorithms.

Co-founder and director at Zeroth Research who develops technical AI-based solutions for customers, with a PhD in mechanical engineering focused on computational mechanics and numerical simulation algorithms.

No summary available yet.

Cindy Wu is an AI safety researcher who participated in the MATS (ML Alignment & Theory Scholars) Winter 2023-24 program (MATS 5.0) and subsequently received funding for an extension period of technical alignment research. She has co-authored two papers in AI safety: "Using Degeneracy in the Loss Landscape for Mechanistic Interpretability" (arXiv:2405.10927, May 2024), which introduces the Interaction Basis technique to address parameter degeneracy that can obscure neural network internal structure, and "Latent Adversarial Training Improves Robustness to Persistent Harmful Behaviors in LLMs" (arXiv:2407.15549, July 2024), where she led experiments on machine unlearning of harmful knowledge. Her research sits at the intersection of mechanistic interpretability and AI safety, with a focus on understanding and improving robustness of large language models. She was affiliated with the MATS program during her research work on both papers.

Develop an LLM-based coordinator and test against consumer spending with 200 people.

6 months of funding for MATS 5.0 extension, with projects on latent adversarial training and persona explainability

Dr Eleanor Drage is a Senior Research Fellow and Assistant Research Professor at the Leverhulme Centre for the Future of Intelligence, where she co‑directs the Narratives and Justice (AI Narratives and Justice) programme. Her research uses feminist and anti‑racist approaches to analyse how AI systems shape power, labour and inequality, and she is co‑investigator on the Desirable Digitalisation project and principal investigator for the HEAT High‑risk EU AI Act Toolkit. She co‑hosts The Good Robot podcast and is author or co‑editor of works including An Experience of the Impossible, Feminist AI and The Good Robot: Why Technology Needs Feminism.

Leonardo Christov-Moore, Ph.D., is a neuroscientist, artist, and complexity theorist who works as a Senior Research Scientist at the Institute for Advanced Consciousness Studies. His research examines trust, empathy, vulnerable AI, mind–body interactions, and the role of non-ordinary and aesthetic experiences in deep belief change, drawing on multiscale neuroscience and neuromodulation methods.

No summary available yet.

No summary available yet.

Intern@Center for Human-Compatible AI, UC Berkeley. Emulating moral progress in AI, to prevent premature value lock-in.

No summary available yet.

A curated directory of 60+ funding sources for AI safety work, maintained by AISafety.com as part of its broader resource hub for the AI safety ecosystem.

UK-based GP, clinical academic, MPhil candidate at University of Cambridge. Founder of GiveHealth, Volunteer Global Health, and Bridge Health Research.

No summary available yet.

A 10-case pilot using independent review and AI accountability to evaluate inscription claims

No summary available yet.

Angelina Wang is an assistant professor of information science at Cornell Tech and in Cornell’s Ann S. Bowers College of Computing and Information Science, where she works on responsible AI. Her research examines the societal impacts of AI, develops methods for evaluating AI systems, and investigates fairness beyond simple, one-size-fits-all mathematical definitions. Previously she was a postdoctoral researcher at Stanford University and completed a Ph.D. in computer science at Princeton University and a B.S. in electrical engineering and computer science at the University of California, Berkeley.

Special Projects Lead at the UK AI Security Institute, working on the societal and ethical implications of advanced artificial intelligence.

3-month stipend and cloud credits to research AI collusion mitigation strategies and develop secure steganography

No summary available yet.

Universal Philosopher, Independent Reasercher

No summary available yet.

Bir ben vardır benden içeri…

Shermon O. Cruz is a Philippine futurist and governance scholar who serves as Executive Director and Chief Futurist of the Center for Engaged Foresight, UNESCO Chair on Anticipatory Governance and Regenerative Cities at Northwestern University (Philippines), and Chair of The Millennium Project’s Philippines Node. His work and teaching focus on futures studies, regenerative cities, and anticipatory governance across government, multilateral, and academic settings.

No summary available yet.

Niels uit de Bos is a mathematician and software engineer working in AI safety and mechanistic interpretability. He holds a PhD in mathematics (algebraic geometry and the geometric Langlands program) from the University of Duisburg-Essen, supervised by Jochen Heinloth, and a Master's degree from Leiden University. He completed a research internship in mechanistic interpretability at MATS 5.0 and 5.1 in 2024, mentored by Adrià Garriga-Alonso at FAR.AI (Foundational AI Research). His primary research output from this period is "Adversarial Circuit Evaluation," a workshop paper at the ICML 2024 Mechanistic Interpretability workshop, which evaluated circuits from the interpretability literature adversarially and found that circuits for the IOI and docstring tasks fail to behave similarly to the full model even on benign inputs. He also co-authored "Relating Piecewise Linear Kolmogorov-Arnold Networks to ReLU Networks," published at AISTATS 2025. Prior to AI research, he had over five years of professional software engineering experience at companies including Zopa, G-Research, and Depict (a Y Combinator ML startup).

No summary available yet.

No summary available yet.