Bundesbank warns of cybersecurity risks from Anthropic's Mythos AI model
Bundesbank President Joachim Nagel warned that Anthropic's advanced Mythos AI model poses cybersecurity risks to financial institutions by identifying and exploiting software vulnerabilities. He called for wider access to the model among institutions to prevent misuse and ensure fair competition, noting that limited access to select US corporations disadvantages financial systems outside the United States.
What happened
Joachim Nagel, Bundesbank president and European Central Bank Governing Council member, warned in a speech on Tuesday that Anthropic's Mythos model poses cybersecurity risks to financial institutions. Nagel stated that Mythos appears capable of quickly identifying and exploiting security vulnerabilities in financial institutions' software.
Nagel called for all institutions to have access to Mythos to prevent misuse and avoid competitive distortions. He emphasized that restricted access to select US corporations creates disadvantages for financial systems outside the United States, including Europe.
Reported impact
Mythos has sparked global fears of a new era of cyber attacks threatening the stability of the financial system worldwide. Regulators, central bankers and corporate executives are seeking to gain more insight on the model, which hasn't been widely released.
Index analysis
The model is described as a double-edged sword that could be used to improve digital security systems or to leverage vulnerabilities for malicious purposes. Nagel's concerns centered on the overall implications of AI on economic growth and price stability.
Governance controls
Nagel called for banking authorities to act to prevent misuse of the technology. His comments suggest regulators view wider access and transparency as a governance approach to managing cybersecurity risks.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.