Bailey Warns Autonomous AI Models Pose Growing Threat

Bank of England Governor Andrew Bailey has sounded the alarm on the risks posed by increasingly autonomous artificial intelligence models. In an article published on the Bank's website, Bailey warned that frontier AI systems are improving themselves through 'recursive learning', which can lead to a self-reinforcing loop where the model governs itself.

According to Bailey, this process reduces society's ability to exercise meaningful oversight and intervention. He emphasized that the greater the capability of the system, the more important it is to address these concerns. Bailey proposed rigorous model testing as a sensible starting point to ensure intervention is both possible and effective.

The Bank of England has previously warned about the potential risks associated with AI investment, particularly in regards to debt financing. The Financial Policy Committee has also highlighted the increasing indebtedness of AI firms, which can complicate risk assessment and amplify losses if expectations disappoint.