Updated · Center for European Policy Analysis · May 21
Anthropic Bars Europe From Mythos Access Over EU AI Act and €15 Million Fine Risk
Updated
Updated · Center for European Policy Analysis · May 21
Anthropic Bars Europe From Mythos Access Over EU AI Act and €15 Million Fine Risk
2 articles · Updated · Center for European Policy Analysis · May 21
Anthropic gave Mythos only to a small group of trusted companies, denying Europe institutional access and even hands-on briefings for the autonomous hacking tool.
EU AI Act rules appear central to that decision: placing a model like Mythos on the European market could trigger systemic-risk scrutiny, technical documentation, adversarial testing and fines of up to €15 million or 3% of global turnover.
Washington’s trust gap with European allies also shaped the exclusion, as frontier cyber models are increasingly treated less like commercial software and more like sensitive national-security technology.
The UK stands apart because its AI Safety Institute already has secure testing capacity and a 2024 US cooperation pact, while the EU has legal authority but no comparable trusted evaluation infrastructure.
That leaves Europe exposed as AI compresses vulnerability discovery from months to weeks or hours, increasing pressure to build secure testing facilities, restricted review pathways and tighter arrangements with the US, UK and Ukraine.
When AI ethics clash with government demands, who controls the nation's most powerful digital weapons?
AI finds security flaws faster than humans can fix them. Is our digital world facing inevitable collapse?
Anthropic’s Mythos AI Under Treasury Review: Systemic Cybersecurity Risks and the Future of Financial Sector Defense
Overview
The U.S. Treasury Department is taking proactive steps to address the urgent risks posed by advanced AI, focusing on Anthropic's Mythos model. Recognizing the rapid advancement of AI's cyber capabilities, Treasury's technology team is seeking direct access to Mythos to identify vulnerabilities before they can be exploited. This move comes as regulators express growing concern about the impact of such powerful AI on the banking sector and financial stability. By engaging directly with Mythos, the Treasury aims to strengthen defenses against evolving cyber threats, highlighting the need for government action as AI becomes increasingly influential in both security and risk.