Resolution is a nonprofit research organization founded in 2026 to pursue a portfolio of theoretical and empirical AI alignment research bets. Its stated aim is to help ensure that advanced AI systems remain beneficial, combining foundational alignment science with heavily automated research intended to accelerate progress. The organization's founders come from the UK AI Security Institute's Alignment Team and from Timaeus, and the group operates under its current name following a rebrand from Sequent / Timaeus Research.
Its research spans learning theory and singular learning theory, scalable oversight, agent foundations, personas, complexity theory, game theory and mathematical autoformalization, alongside theoretical and empirical AI safety work more broadly. Alongside the research portfolio, Resolution builds automated experimentation infrastructure and develops AI research agents, which form the basis of its heavily automated research approach.
The organization also supports external mathematical and formalization projects through grants, including the Formal Commons Project. Resolution has its origins in the United Kingdom and is funded through philanthropy, with major philanthropic funding and API-credit support announced for alignment research; founding researchers have separately led £30 million in Alignment Project funding.
