Alignment Project cohort

The Alignment Project is a global research fund operated by the UK AI Security Institute (AISI). It launched in July 2025 with £15M from AISI and grew to £27M (approximately $35M) by its first-round announcement, with awards going to over 60 research projects and individual grants capped at £1M each. A second round is expected to reopen in summer 2026.

The funder coalition is the distinctive feature. Initial backers included the Canadian AI Safety Institute, CIFAR, Schmidt Sciences, AWS, Anthropic, Halcyon Futures, the Safe AI Fund, UKRI, and ARIA. OpenAI joined later with £5.6M, followed by Microsoft, the Australian Department of Industry's AI Safety Institute, the AI Safety Tactical Opportunities Fund, Sympatico Ventures, and Renaissance Philanthropy. The expert advisory board includes Yoshua Bengio, Zico Kolter, Shafi Goldwasser, and Andrea Lincoln. Routing OpenAI, Microsoft, Anthropic, and AWS money through a UK-government vehicle is uncommon at this scale.

Research areas span interpretability, scalable oversight, alignment theory, and AI control. The control-specific subprogram (documented separately) includes information theory and cryptography (proving theoretical limits on what AI systems can hide or prove about their behavior) and computational complexity theory (formal guarantees and impossibility results behind scalable oversight protocols). This formal-methods framing is unusual for safety grant programs and pulls the cohort toward provable rather than purely empirical work.

The £1M individual ceiling sits between the typical academic grant tier and the larger cohort-style awards from U.S. private funders. It allows the program to fund both small exploratory work and multi-researcher efforts inside the same vehicle. AISI staff manage the fund; the grants are explicitly aimed at the external alignment research community rather than internal AISI projects.

As a UK-government-backed program with international applicants and an industry-plus-philanthropic funder base, the Alignment Project is the largest public commitment to alignment-specific research outside the United States in 2025. Its cross-funder structure is structurally closer to the Frontier Model Forum's AISF than to a single-funder RFP.