The European Central Bank has called the region's largest banks to a Tuesday meeting over cybersecurity risks tied to advanced artificial intelligence models, including Anthropic's Claude Mythos.
ECB Summons Banks Over Mythos Threat
The regulator wants lenders to deploy software patches far faster, according to coverage that first reported the meeting plan. Frank Elderson, vice-chair of the ECB's supervisory board, said years of cybersecurity guidance still hold, but the pace of AI progress now demands quicker action.
Anthropic released the Claude Mythos Preview in April under Project Glasswing, a restricted access program. The model can find unknown flaws in IT systems, and the company says it surfaced thousands of severe vulnerabilities across major operating systems and browsers.
The UK's AI Security Institute found Mythos Preview cleared 73% of expert-level Capture the Flag challenges. No model had passed that benchmark before April 2025.
Elderson said attackers can now reverse-engineer a fix within 30 minutes, so the slower update cycles common at many banks no longer suffice.
Also Read: XRP Eyes $1.50 Breakout As Exchange Supply Tightens
Why Regulators See Urgency
The ECB supervises 111 of the largest banks in the Eurozone, and most of them sit outside Project Glasswing. That leaves European lenders without direct access to frontier models like Mythos, a gap Elderson called unfortunate.
He wants US institutions at Tuesday's session to share testing insights with their European counterparts.
Lack of access cannot justify inaction, Elderson argued, because malicious actors could soon reach the same technology. He warned that "andante" tempo is no longer enough, and that supervisors now need banks moving at "presto" speed.
The pressure is not confined to Europe. The European Commission is negotiating with Anthropic over testing companies and banks for vulnerabilities that Mythos uncovers, while French startup Mistral AI is in talks to offer European banks its own flaw-hunting tool.
What the Mythos Rollout Has Shown
Elderson's warning caps weeks of escalating regulatory attention since Mythos reached a small set of US banks. In mid-May, he urged Eurozone lenders to prepare for AI-assisted attacks, telling the ECB's Supervision Newsletter that the access gap made the threat more severe rather than less. Wall Street watchdogs paused some cyber examinations after the model exposed unexpected weaknesses, and Mozilla shipped Firefox 150 with 271 patches for bugs the model identified. Japan's three largest banks are expected to gain access in the coming weeks, widening a divide that European supervisors now appear determined to close.
Read Next: Bitcoin Rally Hits A Ceiling As Sellers Guard $77,050 Resistance