The European Central Bank has summoned major banks to an emergency meeting to warn of new cybersecurity risks linked to advanced Artificial Intelligence models. The regulator said the latest models could reshape the threat landscape for banks by accelerating how quickly attackers can discover and exploit weaknesses in digital systems.
Frank Elderson, vice chair of the ECB’s Supervisory Board, said banks must become significantly faster at installing security updates. According to the ECB, new Artificial Intelligence tools can make it possible to identify and exploit security flaws much faster than in the past. The warning reflects concern that defensive practices in banking are not moving quickly enough to match the pace of new capabilities.
One example is Anthropic’s Artificial Intelligence model, Claude Mythos Preview. According to the company, the model has detected thousands of serious vulnerabilities in operating systems and web browsers. That ability has heightened concern that similar tools could give malicious actors a faster path to uncovering exploitable flaws across widely used software.
European banks are considered more vulnerable because many lack access to the technology, while some US banks are already testing it. The ECB is hoping US banks will share their experiences with European competitors. The gap in access and experimentation has raised fears that parts of Europe’s banking sector could be less prepared to respond as advanced Artificial Intelligence systems become more deeply involved in cybersecurity.
