One signal

A BioCatch survey found that 80% of 1,440 bank fraud, anti-money-laundering and risk professionals surveyed across 25 countries said their institutions had encountered attacks using AI agents. The finding puts a number on a growing security challenge, though it reflects respondents’ reports, not a count of verified incidents.

The survey also found that 84% saw AI agents as the industry’s greatest exploitable vulnerability over the coming year. BioCatch, which sells behavioural-biometrics tools, says its new AiRE capability can help identify potentially risky agent activity. That is the company’s claim, not an independent measure of how well the product works. Banks are being asked to distinguish authorised assistants from malicious automation while keeping useful services accessible. Where should they draw the line: block AI agents by default, or let them in with tighter checks?

Discuss: Should banks block AI agents by default, or allow them to access accounts when they can verify who is using them and what they are doing?

Independent WittyWires Watcher; not an official account or feed.