London: Mythos AI cyber risks will come before the Financial Stability Board after Anthropic agreed to brief the global watchdog on flaws found in financial systems, the Financial Times reported.
Bank of England Governor Andrew Bailey requested the briefing. Bailey chairs the Financial Stability Board, which coordinates financial rules and risk monitoring for G20 economies.
Anthropic will discuss Mythos Preview, its unreleased cybersecurity model. The company says the model can detect old flaws in web browsers, infrastructure, and software.
Cybersecurity experts warned that the same tool could help attackers build more advanced cyberattacks. Banks face added risk because many still rely on older systems that take time to patch.
Bailey warned about the model in April at Columbia University in New York. He said Anthropic may have found a way to “crack the whole cyber risk world open.”
He also questioned whether the model could identify weaknesses that others might exploit. His remarks placed advanced AI tools inside the wider financial-stability debate.
Mythos can help defenders find flaws faster. However, it also says the model could compress attack timelines and widen the gap between discovery and patching.
Restricted access through Project Glasswing. It says vetted groups, including Apple, Microsoft, and JPMorgan Chase, have used the model to scan and secure systems.
Read: Anthropic Launches Project Glasswing to Fight AI Cyber Threats
Reuters said it could not immediately verify the Financial Times report. Anthropic and the Financial Stability Board did not immediately respond to Reuters requests for comment.
Mythos AI cyber risks now sit in a wider regulatory debate over dual-use AI. Regulators are weighing whether tools built for defence could also pose systemic threats to global finance.
Anthropic is set to brief the Financial Stability Board on cyber vulnerabilities in global finance found by its unreleased Mythos Preview AI model.
