Regulators in five countries monitoring Anthropic's Mythos due to banking risks

The parade of regulators has grown: Australia's Securities and Investments Commission (ASIC) has officially confirmed that it is monitoring the development of Mythos — a new language model from Anthropic. This comes against the backdrop of escalating international reaction: over several weeks, isolated concerns have transformed into coordinated global monitoring. Before ASIC, the Bank of England, the U.

S. Federal Reserve, and the U.S.

Treasury Department had already announced they were monitoring Mythos. ECB President Christine Lagarde added a critical warning to the general concern: no management system for the new model exists yet. For the eurozone's financial architecture with a GDP exceeding 14 trillion euros — these are not merely words.

What is Mythos and why has it sparked such a reaction? Anthropic — a company known primarily for its family of Claude models — is developing Mythos as a system with enhanced autonomous action capabilities in the financial environment. According to available information, the model can interact with banking infrastructure, process transactions, and make financial decisions with minimal human involvement.

It is precisely this that has drawn the attention of regulatory bodies: autonomous AI in the financial sector represents a fundamentally different level of risk compared to conventional analytical tools. Regulators are concerned about several scenarios. The first is systemic risk: if a powerful AI model becomes widely deployed in the banking sector and makes an error, the consequences could spread across the entire market — many institutions would operate according to the same algorithm.

The second is lack of transparency: the principles underlying large language models are difficult to verify, which means regulators cannot guarantee compliance with established rules. The third is the pace of change: AI companies release new versions faster than standards can be formed. Anthropic positions itself as a company prioritizing safety.

The philosophy of constitutional AI and public commitments to responsible development are part of its corporate identity. Nevertheless, regulators' response demonstrates: declarations of safety and actual readiness of the system to operate in critical infrastructure are different things. Oversight bodies want to see concrete mechanisms for audit, control, and emergency shutdown.

What is happening fits into a broader trend. Following the rapid growth of 2023–2024, when AI tools began penetrating financial services, regulators worldwide are transitioning from observation to active measures. In Europe, the AI Act is in force, establishing requirements for high-risk systems.

In the U.S., Congress is working on legislation, while the SEC, CFTC, and OCC are formulating their own approaches.

Australia traditionally follows international standards of financial regulation — and, judging by ASIC's statement, has no intention of falling behind. For companies developing AI for the financial sector, a new reality is arriving: a model's capabilities no longer determine the horizon of its application. Regulatory acceptability is now as much a constraint as technical readiness.

Mythos can be as powerful a tool as possible, but without clear frameworks of accountability and coordination with regulatory bodies, its commercial deployment in the banking sphere remains in question. For Anthropic, this is a moment of reckoning. The company has built its reputation on the principle of safety first — now it must prove that these principles work not only in the laboratory, but also under the scrutiny of five of the world's largest financial regulators.