UK AI Security Institute ICLR 2026 Social

​Join us to chat about research, policy, and how to make advanced AI go well.

​Join us for food and drinks at our lovely venue, a 5-min drive (or free ICLR shuttle) from Riocentro.

​About AISI:

​We're a fast-paced research organisation inside the UK government. We're working to develop the deepest technical understanding of frontier AI's risks and what to do about them, then galvanize the UK, international governments, and developers to mitigate them. We have about 100 technical staff, many of whom have joined from places like Google DeepMind, OpenAI, and Anthropic.

​Our work includes:

• ​Frontier research on technical solutions for safeguards, control, and alignment

• ​Pre-deployment evaluations of models such as OpenAI's o1, Anthropic's Claude 3.5 Sonnet, and Claude Mythos

• ​Inspect, our open-source evaluation framework, and the new Inspect Cyber

​We're scaling fast and have lots to do (we're hiring for research scientist, research engineer, and other technical roles). Learn more about our plans at aisi.gov.uk/research-agenda.

​Please be aware that photographs will be taken at the event. By attending the event, you consent to us taking your photo and storing it for marketing purposes. You also affirm that you are over 21 years of age.