The UK AI Security Institute (AISI) works closely with its international counterparts and the UK’s closest security partners to share the latest information on frontier AI model capabilities and risks, especially as they relate to national security matters.

UK AISI has formal partnership agreements with its counterparts in Singapore, Australia, and the United States. These agreements allow for advanced bilateral information sharing on AI model capabilities and secondments of technical talent, as well as enabling the Institutes to conduct joint studies into AI risks.