OpenAI and Microsoft have pledged funding to the UK’s AI Security Institute (AISI) to support its flagship Alignment Project, an international initiative focused on ensuring that advanced AI systems are safe, secure, and behave as intended. The announcement, made at the AI Impact Summit, brings the total funding for the project to over £27 million, including £5.6 million from OpenAI and additional contributions from Microsoft and other partners.
The Alignment Project is designed to foster trust in AI technologies as they are increasingly adopted in public services, healthcare, and other sectors. Its first round of grants has already supported 60 projects across eight countries, with a second round scheduled for later in 2026. Alignment research focuses on preventing unsafe or unintended behaviours in AI systems as they grow more capable, promoting reliability, safety, and public confidence.
UK Deputy Prime Minister David Lammy highlighted that the initiative ensures AI safety is integrated from the outset, enabling the country to realise the benefits of AI technologies responsibly. AI Minister Kanishka Narayan emphasized that public trust is crucial for widespread adoption, making alignment research a priority.
The project benefits from a broad international coalition of backers, including research institutions, government agencies, tech companies, and philanthropic organisations such as CIFAR, Anthropic, AWS, Schmidt Sciences, and UKRI. AISI also provides academic mentorship, grant funding, and access to compute resources, creating a collaborative global ecosystem for AI safety research.
Mia Glaese, VP of Research at OpenAI, noted that alignment challenges require diverse, independent approaches and that supporting the Alignment Project complements OpenAI’s internal research. The initiative cements the UK’s role as a leader in global AI safety efforts, leveraging its world-class universities, research institutions, and technology companies to drive progress in developing predictable, controllable AI systems.
