Whalesbook
Global Scrutiny on Mythos AI
Banks worldwide are rushing to adopt Anthropic's advanced AI model, Mythos. This rush comes as regulators globally step up their review of its serious cybersecurity risks. Experts recognize the AI model for its strong cyber offensive capabilities, causing significant worry among financial regulators.
Restricted Access Spurs Race
Access to Mythos has been tightly limited, with only a select few U.S. institutions like JPMorgan and Bank of America gaining early entry. This scarcity fuels fierce competition and a wider industry race to keep pace. Deutsche Bank CEO Christian Sewing noted the need for careful risk management, calling it a situation requiring constant vigilance rather than immediate panic. He confirmed close contact with European regulators.
Major Banks Confirm Use
Other major U.S. banks have also confirmed they have access to Mythos. Morgan Stanley CEO Ted Pick mentioned his bank's permission for the 'Claude Mythos Preview,' pointing to the growing cyber threat landscape. Goldman Sachs CEO David Solomon said his firm has the model and is working with Anthropic and security partners to use its advanced features. Citigroup has also been running internal tests with the AI technology.
Global Warnings Issued
The UK government issued a stark warning to Anthropic in an open letter. Its AI Security Institute found Mythos to be "substantially more capable at cyber offence than any model we have previously assessed." Asian regulators are watching developments closely, with South Korea's Financial Supervisory Service reviewing potential risks. European supervisors are checking how banks are ready through their existing cyber security measures, aiming to understand awareness and response plans. Barclays CEO C. S. Venkatakrishnan called Mythos a serious threat, expecting similar, more powerful cyber threats to follow.