Bundesbank Urges EU to Secure Anthropic's Mythos AI Access for Bank Cybersecurity
European Banks Face Cybersecurity Challenges Amid Restricted Access to Mythos AI
By Reinhard Becker
Anthropic's Mythos AI and Its Impact on Banking Cybersecurity
BERLIN, April 29 (Reuters) - European banks need to be given access to Anthropic's latest artificial intelligence model, Mythos, if they are to shield themselves against the threat of cyberattacks powered by this new breed of programmes, one of Germany's top financial regulators told Reuters.
Mythos, a model designed to find flaws in computer code, is viewed by cybersecurity experts as potentially turbo-charging attacks on banks' technology systems but so far it has only been made available to some U.S. banks.
Bundesbank's Call for EU Intervention
The chief supervisor at Germany's Bundesbank, Michael Theurer, urged European Union authorities to request access to Mythos from Anthropic or from Donald Trump's U.S. administration.
"I consider it necessary that the European Commission and governments in Europe now also approach the company — or rather the United States — to request that the technology be shared," he told Reuters in an interview. "There has to be an official request so that we in Europe can also benefit from the insights."
Potential for Wider Access in Europe
Sources recently told Reuters Anthropic plans to provide access to its Mythos AI model to European banks soon.
Regulatory and Industry Response to Mythos AI
Concerns Raised by Mythos' Limited Release
Mythos' release to a selected group of companies, initially including just one bank, earlier this month raised alarm among regulators and policymakers across the world.
European Banks' Preparedness and Ongoing Assessments
German banks and national authorities have been jointly examining risks around these models and European Central Bank supervisors, who are responsible for the euro zone's biggest lenders, have been quizzing bankers about their preparedness.
Theurer said his meetings with banks so far showed they were acutely aware of the possible threat from the new model but they could not "test which vulnerabilities it is capable of identifying" without getting access to it.
Risks of Technological Dependence
"We may be moving into an area in which economic actors could potentially become dependent on state assistance," Theurer said. "If you do not have access to this technology yourself, you naturally find yourself in very difficult waters."
Regulators' Capacity to Address AI Risks
Regulators' own ability to monitor and combat the risks posed by Mythos and similar models has been called into question, with a survey finding authorities significantly lag financial firms in AI adoption and lack data on emerging harms.
(Writing by Francesco Canepa in Frankfurt; Editing by Toby Chopra)
