The Bank of England and the FCA published the results of their survey on artificial intelligence in UK financial services on 26 October, based on 196 responding firms. 86% already use AI, and generative AI accounts for 56% of systems. 80% have adopted AI agents, but only 2% give them broad autonomy. Third-party systems make up 39% of the total and 53% of high-materiality systems, while the three largest cloud providers account for 79% of relationships.

What it means: The figures show that artificial intelligence is already an everyday tool in banks, insurers and investment firms, but the firms themselves are wary of letting AI agents decide on their own. The most important finding for supervisors is concentration: when over half of high-materiality systems come from third parties and nearly four in five cloud relationships involve three providers, an outage or security gap at one provider can hit many institutions at once. The survey also records that a large majority of firms expect frontier models to affect cyber security and that many have already stepped up attention to patching. In the EU the same issues are covered by DORA, which requires a register of ICT provider contracts and oversight of critical third-party providers, and by the AI Act. For Greek banks, payment institutions and fintechs, the UK survey is a useful benchmark: they should map how many AI systems they use, from which providers and with what level of human oversight, before the Bank of Greece asks them to.

Read the original on Bank of England / FCA

Related service: myDATA integrations

Related stories