From narrow AI to superintelligence
Foster the considered, safe development of artificial intelligence — from narrow AI to general and superintelligence — in the interest of all humanity.
Two vectors, eight directions
Interpretability
Understanding how and why models make decisions.
Value alignment
Bringing AI goals into line with human values.
Common-sense synthesis
Giving systems a robust understanding of the world and causality.
Energy efficiency
Reducing the compute and energy cost of AI.
Lifelong education
Preparing people to work and live alongside AI.
Business & public-sector transformation
Responsible adoption of AI in processes and services.
Updating legislation
Legal frameworks for the safe development of AI.
Digital equality
Equal access to the opportunities of AI.
Where the dialogue continues
AI Forum Review
Latest news
The Forum launches an interpretability programme
The new programme will unite researchers and practitioners around model-explainability and audit methods.
New members of the Scientific Committee
Researchers in alignment and AI law have joined the committee.
Results of the energy-efficient AI hackathon
Teams presented solutions reducing model-inference energy use by up to 40%.