The European Central Bank (ECB) has summoned banks to a meeting this week to push them to urgently address IT and cybersecurity vulnerabilities exposed by Anthropic’s powerful new Mythos AI model.
Frank Elderson, vice-chair of the ECB supervisory board, warned that while many cyber issues have been known for years, rapid advancements in AI demand much faster action. “The clock is ticking,” he said.
Elderson highlighted that once software patches are released, malicious actors can reverse-engineer vulnerabilities in as little as 30 minutes. Banks must therefore significantly speed up their patching processes beyond current market practices.
Anthropic has denied European banks and regulators access to Mythos. While a small group of large US banks with access have already identified hundreds to thousands of low-to-moderate vulnerabilities — some of which can be chained into major threats — the ECB is urging them to share critical findings with their European counterparts.
However, Elderson stressed that lack of direct access to the model is no excuse for inaction. “Malicious actors might have access to this technology soon,” he cautioned.
The ECB’s message is clear: banks must treat the risks posed by advanced AI tools like Mythos with utmost seriousness and act immediately to strengthen their defenses.
