The AI Safety Institute is the first state-backed organisation dedicated to advancing this goal.
We are conducting research and building infrastructure to test the safety of advanced AI and to measure its impacts on people and society. We are also working with the wider research community, AI developers and other governments to affect how AI is developed and to shape global policymaking on this issue.
As a complement to our empirical evaluations of frontier AI models, AISI is planning a series of collaborations and research projects sketching safety cases for more advanced models than exist today, focusing on risks from loss of control and autonomy. By a safety case, we mean a structured argument that an AI system is safe within a particular training or deployment context.
We tested leading AI models for cyber, chemical, biological, and agent capabilities and safeguards effectiveness. Our first technical blog post shares a snapshot of our methods and results.
Since February, we released our first technical blog post, published the International Scientific Report on the Safety of Advanced AI, open-sourced our testing platform Inspect, announced our San Francisco office, announced a partnership with the Canadian AI Safety Institute, grew our technical team to >30 researchers and appointed Jade Leung as our Chief Technology Officer.
Monitoring the fast-moving landscape of AI development
Evaluating the risks AI poses to national security and public welfare
Advancing the field of systemic safety to improve societal resilience
Working with AI developers to ensure responsible development
Informing policymakers about current and emerging risks from AI
Promoting global coordination on AI governance
For our ambitious and urgent mission, we need top talent. We have built a unique structure within the government so we can operate like a startup. We have recruited over 30 technical staff, including senior alumni from OpenAI, Google DeepMind and the University of Oxford, and we are scaling rapidly. Our staff are supported by substantial funding and computing resources, priority access to top models, partnerships with leading research organisations and an incredibly talented, close-knit and driven team.
