Research and analysis
Frontier AI: capabilities and risks – discussion paper
A discussion paper on the capabilities of, and risks from, frontier AI.
Documents
Details
This discussion paper was written to inform discussions at the AI Safety Summit 2023. It does not represent government policy.
A shared understanding of the risks is essential to unlock the enormous benefits of frontier AI. This paper outlines the capabilities, risks, and cross-cutting challenges presented by the technology, pointing particularly to dangers around misuse, social harms, and loss of control.
The paper was reviewed by an expert panel composed of external experts, including Turing Prize Winner Yoshua Bengio, and other leading minds such as Sara Hooker, Arvind Narayanan, William Isaac, Paul Christiano, Irene Solaiman, Alexander Babuta and John McDermid.
Updates to this page
Published 25 October 2023