Bank of England Governor Says AI Regulation Is Not the Right Place to Start

Bank of England Governor Andrew Bailey says regulating artificial intelligence (AI) is not the right place to start. He argues that rigorous testing to identify vulnerabilities and safeguards to contain risks should come first.

In his first article for Substack, Bailey described AI risks as real and increasingly significant. He said development should not be halted or prohibited, citing the immense benefits, but called for a system to intervene and set boundaries for how AI operates.

Bailey said important work is underway in the UK, including through the AI Security Institute, and that testing should help form part of a set of standards. He called for the pace of progress to accelerate, while urging humility. Models may fail or behave unexpectedly during testing, he said; that does not show testing has failed, but demonstrates why it is necessary.

Bailey said testing should not replace future regulation. A more formal regulatory framework might emerge over time, he said, but regulation is not where he believes the process should begin.