A new artificial intelligence system developed by Anthropic is raising alarm among cybersecurity experts, who warn it could significantly increase cyberattack risks for global banking systems.
The model, called Claude Mythos Preview, was announced earlier this month and is described by the company as its most advanced system yet for coding and autonomous “agent-like” tasks.
Security researchers say the tool’s ability to write and analyze code at a high level could also allow it to discover and exploit hidden vulnerabilities in complex digital systems, including those used by financial institutions.
Experts warn this could be especially dangerous for banks such as JPMorgan Chase, which rely on a mix of modern software and decades-old legacy systems that are harder to secure.
Cybersecurity specialists say banks are particularly exposed because they often use similar third-party systems for critical operations like customer verification, payments, and identity checks. This could allow a single exploit to spread quickly across multiple institutions.
A Cloud Security Alliance briefing said Mythos represents a “step change” in AI capability, lowering the skill level needed to discover and exploit software weaknesses before companies can patch them.
According to Anthropic researchers, the model has already identified thousands of high-risk security flaws during testing, including long-standing vulnerabilities in widely used software tools.
Officials from the United States, United Kingdom, and Canada have reportedly met with banking representatives to discuss potential risks linked to advanced AI-driven cyber threats.
The model is not being released publicly. Instead, Anthropic has launched a controlled evaluation program involving selected technology companies and financial institutions to test its defensive uses.
Some experts say the development highlights a growing challenge for the financial sector: defending increasingly complex systems against AI tools that can also be used for offensive cyber operations.
Industry voices warn that without stronger cybersecurity upgrades, AI-powered tools like Mythos could significantly raise the scale and speed of future cyberattacks on global financial infrastructure.
0 Comments