Global regulators scramble as AI model raises banking cyber risk fears

Yesterday 14:40

By: Dakir Madiha

Global regulators scramble as AI model raises banking cyber risk fears

Financial authorities across Europe, the United Kingdom, and the United States have begun coordinated assessments of cyber risks linked to a powerful new artificial intelligence model developed by Anthropic. The system, known as Mythos Preview, has demonstrated the ability to identify and exploit previously unknown software vulnerabilities at a scale and speed that has alarmed regulators and banks.

Supervisors at the European Central Bank are preparing to question banks about their readiness to handle emerging cyber threats tied to the model. In Germany, the BaFin and the Bundesbank are consulting with financial institutions and cybersecurity experts to assess exposure. Authorities warned that firms must prepare for the near term discovery of vulnerabilities requiring rapid mitigation.

In the United Kingdom, the Bank of England, the Financial Conduct Authority, and HM Treasury are coordinating with national cybersecurity agencies to evaluate risks. The issue has been added to key financial stability forums, with briefings for major banks and market infrastructure operators expected within weeks. In the United States, Treasury Secretary Scott Bessent and Federal Reserve Chair Jerome Powell have already urged major financial institutions to test their systems against the model’s capabilities.

Anthropic has restricted access to Mythos Preview through a controlled initiative known as Project Glasswing. The program includes major technology and finance partners such as Amazon Web Services, Microsoft, Google, Apple, JPMorgan Chase, and Nvidia. The company has committed substantial funding to support security testing and open source defenses, while limiting access to organizations managing critical infrastructure.

Major Wall Street banks including Goldman Sachs, Citigroup, Bank of America, and Morgan Stanley have begun internal testing of the model. Meanwhile, cryptocurrency platforms Coinbase and Binance are seeking access as they evaluate how the technology could affect digital asset security. Industry experts warn that the model’s ability to chain together zero day vulnerabilities dramatically shortens the time between discovery and exploitation, compressing what once took months into minutes.

Security specialists say the development represents a structural shift in cyber risk. The acceleration of both offensive and defensive capabilities is forcing institutions to rethink how they monitor systems, patch vulnerabilities, and respond to threats in real time. Regulators are now racing to understand whether existing safeguards are sufficient for a new era shaped by autonomous AI driven cyber operations.