US and European regulators question banks about AI agents and cyber risks

Banking regulators in the United States and Europe are stepping up efforts to include the dangers AI postures to banks, Reuters reportedmentioning 6 individuals knowledgeable about the matter. Some fear that cyberattacks or rogue AI representatives might shake the monetary system.

In the United States, the Federal Reserve and 2 other bank regulators are questioning banks about how they utilize sophisticated AI, 2 sources stated.

They need to know if banks utilize Chinese-built tech, usage AI representatives for cyber defence, and can manage systems that act upon their own.

Regulators likewise wish to know what limitations banks place on AI representatives, whether human beings still make crucial choices, and how banks would shut those systems down. All 3 United States companies decreased to comment to Reuters.

In Europe, the Bank of England and the European Central Bank are pressing banks to enhance their prepare for recuperating from attacks, 3 individuals stated.

Both fret that quick development in AI might trigger severe disturbance in monetary markets earlier than anticipated.

UK managers desire the biggest banks to evaluate backup systems that are totally different from their primary IT systems, 2 sources stated.

These can keep a bank running if its primary systems are hacked, and the UK would rather deal with banks than set brand-new guidelines, they included.

The ECB provided loan providers up until completion of October to produce strategies versus AI-enabled cyber hazards, Reuters reported. Eurozone banks are likewise being asked to map where they utilize AI, consisting of through outdoors providers, so they can change systems off and reboot them securely.

“The adoption of AI is frequently outmatching the capability to put the controls in location,” Michael Woodbridge, a cybersecurity partner at Grant Thornton, informed Reuters.

The examination follows a run of AI-linked security occurrences. Previously this month, security company CrowdStrike stated an individual in China likely utilized a Chinese-built AI representative and Anthropic’s Claude Code to hack South Korean banks. OpenAI’s rogue representatives likewise burglarized Hugging Face previously this year.

On Tuesday, the International Monetary Fund likewise cautioned that a significant cyberattack might harm self-confidence in monetary markets, in a stability report that flagged dangers from AI.