The Alignment Project↗

No longer funding AI safety
All 56 grants ↓
Funder·Founded 2025·London, UK
“Funding groundbreaking AI alignment research to keep advanced AI systems safe, secure and beneficial to society.”In its own words: alignmentproject.aisi.gov.uk
Technical

Pooled alignment-research fund run by the UK AI Security Institute with government, industry and philanthropic partners; its first round awarded over £27m to more than 60 projects, and applications are currently closed.

Show detailsHide details

A pooled global fund of over £27 million, hosted by the UK AI Security Institute, that funds alignment research with grants of up to £1 million per project. An expert advisory board guides its direction. Its first round awarded over £27 million to over 60 projects.

Typical grant $590k

Itemised here $33M in 56 grants

Checked 29 Oct 2026

Money flows

Where its money came from and where it went. Pick years and what to show; hover a box or ribbon for details, click one to focus it, click the background to see everything.

How to apply

Applications
Closed
AISI says it is unlikely to run again in 2026; other academic grant programmes expected via aisi.gov.uk/grants
Who can apply
Academics, For-profits, Individuals, New organisations, Organisations

Biggest recipients

Money from

Gifts and pooled money between funders, as recorded on the site.

All grants

56 grants totalling $33M, newest first. Recipients link to their own pages; people who received grants without a public profile are shown by number.

2026 56 grants · $33M
DateRecipientAmountPurpose
19 Feb 2026ELLIS Institute Tübingen~$590kAlignment Project round 1 (lead: Abdelnabi, Sahar): Foundations of Situational Awareness Evaluation and Control. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026New York University~$590kAlignment Project round 1 (lead: Individual #34065): What Can Mr. Hyde Tell Us About Dr. Jekyll? Efficiently Evaluating and Red-teaming "Safe" Models with Unsafe Alter Egos. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026LawZero~$590kAlignment Project round 1 (lead: Bengio, Yoshua): Designing a safety case and latent variable model for the non-agentic Scientist AI. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Yale University~$590kAlignment Project round 1 (lead: Bergemann, Dirk; Morris, Stephen): Joint alignment and control via mechanism and information design. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Stanford University~$590kAlignment Project round 1 (lead: Silverstein, Eva): Precision optimization and symmetry breaking for improving AI predictability. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Anti Entropy~$590kAlignment Project round 1 (lead: Brown, Jason): Investigating the potential for AI debate to autonomously provide meaningful oversight on difficult-to-evaluate safety-relevant tasks. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University of Warwick~$590kAlignment Project round 1 (lead: Condorelli, Daniele): How does AI learn to play games? Theory and experiments. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Flourishing Future Foundation~$590kAlignment Project round 1 (lead: De Lucena, Diogo): When Models Refuse to Be Steered: Understanding Self-Monitoring and Alignment Drives. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Stanford University~$590kAlignment Project round 1 (lead: Duchi, John): (Mis)aligned preferences: what data is needed for agents to reflect honest human preferences. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Ashgro Inc~$590kAlignment Project round 1 (lead: Individual #43700): Training dynamics theory for AI alignment through early detection of phase transitions and stagewise learning dynamics. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Imperial College London~$590kAlignment Project round 1 (lead: Duncan, Andrew): PRISM: Probabilistic Rare-event Inference for Safety of Models. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Ashgro Inc~$590kAlignment Project round 1 (lead: Gietelink Oldenziel, Alexander): Loss Landscape and Training Dynamics for (ReLU) Neural Networks. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University of California, Berkeley~$590kAlignment Project round 1 (lead: Guruswami, Venkatesan): Error-correction and Pseudorandomness in Alignment: Watermarks and Backdoors. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Alignment Research Center~$590kAlignment Project round 1 (lead: Individual #4327): Low probability estimation: mechanistically estimating the probability of rare failures in machine learning systems. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University of Chicago~$590kAlignment Project round 1 (lead: Individual #3276): Hyperstition in LLMs: how narratives become model policies, their persistence after removal, and tamper-evident tripwires. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Timaeus Research~$590kAlignment Project round 1 (lead: Hoogland, Jesse): Patterning: Singular Learning Theory for Alignment. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University of Liverpool~$590kAlignment Project round 1 (lead: Huang, Xiaowei): Rare-Event Estimation in Large Language Models via Subset Simulation and Contrastive Unalignment. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Principles of Intelligence~$590kAlignment Project round 1 (lead: Individual #9498): Constructing robust incentives for AI agents to act cautiously. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Yale University and University of Michigan~$590kAlignment Project round 1 (lead: Cai, Yang; Haghpanah, Nima; Immorlica, Nicole; Liponowski, Elliot; Ravid, Doron): Is the AI Misaligned or Misinformed? Applying and further developing the strategic communication framework. Amount undisclosed; awards £50k–£1M from a £27m pool acro source
19 Feb 2026New York University~$590kAlignment Project round 1 (lead: Izmailov, Pavel): Weak-to-Strong Consistency Reinforcement Learning. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University of Alberta~$590kAlignment Project round 1 (lead: Jiang, Bei): Adaptive Tilting to Quantify Tail Risks in Aligned Language Models. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University of Toronto~$590kAlignment Project round 1 (lead: Jin, Zhijing): Formal verification and Empirical Testing of Translucency in Multi-Agent AI Systems. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University of Toronto~$590kAlignment Project round 1 (lead: Jin, Zhijing): Game-Theoretic Safety Guarantees for Advanced AI Systems. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Recursive Safeguarding Ltd~$590kAlignment Project round 1 (lead: Kaddar, Younesse): Bayesian Scientist AI for LLM Agent Alignment: GFlowNet-sampled plan abstractions and white-box monitors to mitigate drift and deception. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University of Toronto~$590kAlignment Project round 1 (lead: Kahn, Yonatan; Levi, Noam; Larkoski, Andrew): Scaling laws, data distributions, and learning dynamics: simulated high-energy physics data as a benchmark. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University of Chicago~$590kAlignment Project round 1 (lead: Kondor, Risi): Probabilistic Agentic Foundations for Mesa-Optimization and Emergent Misalignment. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University of Alberta~$590kAlignment Project round 1 (lead: Kong, Linglong): Sample-Efficient Online Fine-Tuning Against Resistant Behaviors: Statistical Foundations for Post-Training Alignment. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026London School of Economics and Political Science~$590kAlignment Project round 1 (lead: Levy, Matthew): Competition and the AI Control Problem: Mechanism Design with Opaque Agents. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Northwestern University~$590kAlignment Project round 1 (lead: Liang, Annie): Friend or Foe: Delegating to an AI whose Alignment is Unknown. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Center for AI Safety~$590kAlignment Project round 1 (lead: Meek, Austin): Steering Models Towards Safer Reasoning. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Purdue University~$590kAlignment Project round 1 (lead: Psomas, Alexandros): Beyond Nash: Pseudo-Equilibria for Robust AI Alignment. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Carnegie Mellon University~$590kAlignment Project round 1 (lead: Raghunathan, Aditi): Enhanced Cyber Control Evaluations: An Empirical and Learning-Theoretic Approach. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Anti Entropy~$590kAlignment Project round 1 (lead: Recchia, Gabriel): Synthetic benchmarks on how argument types and evidence structures affect judge accuracy in debate-based AI oversight. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Anti Entropy~$590kAlignment Project round 1 (lead: Recchia, Gabriel): How debate assistance, annotator expertise and time constraints affect human feedback accuracy. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Arcadia Impact~$590kAlignment Project round 1 (lead: Individual #10205): Building alignment research capacity in collaboration with AISI alignment team. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Ashgro Inc~$590kAlignment Project round 1 (lead: Rosas, Fernando): Cybernetic Foundations for Aligned Agency: From Internal Models to Verifiable Control. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University of Pennsylvania~$590kAlignment Project round 1 (lead: Saeedi Bidokhti, Shirin): Strategic Behavior and Manipulation in Large Language Models: Experiments, Theory, and Regulatory Mechanisms. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University College London and University of Oxford~$590kAlignment Project round 1 (lead: Individual #9720): Alignment Training Dynamics. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University of Oxford~$590kAlignment Project round 1 (lead: Santhanam, Rahul): A Theoretical Framework for Machine Understanding. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University College London~$590kAlignment Project round 1 (lead: Saxe, Andrew): Theoretical Model Organisms of Misalignment: Learning Dynamics of Misalignment and Theory-Grounded Mitigation. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Columbia University~$590kAlignment Project round 1 (lead: Servedio, Rocco; Chen, Xi; De, Anindya): Discovering rare harmful behaviors exhibited by high-dimensional AI systems. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Simplex~$590kAlignment Project round 1 (lead: Shai, Adam): Simplex: A Principled Science of Interpretability. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Toulouse School of Economics~$590kAlignment Project round 1 (lead: Smolin, Alex; Dworczak, Piotr): Designing Human-AI Interactions: Misalignment, Robustness, and Trust. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026UC Berkeley Center for Responsible, Decentralized Intelligence~$590kAlignment Project round 1 (lead: Song, Dawn): Model Organisms of Reward Hacking and Long-Horizon Deception for Testing RL Alignment. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Tel Aviv University~$590kAlignment Project round 1 (lead: Talgam-Cohen, Inbal): Mitigating Moral Hazard in AI using Contracts. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Independent Researcher~$590kAlignment Project round 1 (lead: Individual #46405): Investigating Dynamics of Agentic Monitoring in AI Control. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Massachusetts Institute of Technology~$590kAlignment Project round 1 (lead: Vaikuntanathan, Vinod): Verification, Robustness and Alignment through a Cryptographic Lens. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Massachusetts Institute of Technology~$590kAlignment Project round 1 (lead: Vaikuntanathan, Vinod): Backdoors and Cryptographic Hardness in Neural Networks. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Stanford University~$590kAlignment Project round 1 (lead: Van Roy, Benjamin): A Mathematical Model of Misalignment. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University of Oxford~$590kAlignment Project round 1 (lead: Individual #43103): Adversarial Elicitation Theory. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026King's College London~$590kAlignment Project round 1 (lead: Ventre, Carmine): The Implementability-Capability Frontier: Mechanism Design for Robust AI Alignment. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Monash University~$590kAlignment Project round 1 (lead: Wei, Susan): Probabilistic Interpretability of Transformers. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Massachusetts Institute of Technology~$590kAlignment Project round 1 (lead: Williams, Ryan): Exploring The Role of Relativization and Problem Extensions in AI Safety. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University of Waterloo~$590kAlignment Project round 1 (lead: Individual #9204): Safe Superintelligence via AIXI Agent Foundations. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026University of California, Berkeley~$590kAlignment Project round 1 (lead: Yu, Bin): The Geometry of Risk: Using Model Complexity to Improve Predictions of Catastrophic AI Failures. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source
19 Feb 2026Tel Aviv University~$590kAlignment Project round 1 (lead: Zamir, Or): Provable and Rigorous Notions of Security for Machine Learning via Cryptography. Amount undisclosed; awards £50k–£1M from a £27m pool across 60 announced projects. source