Anthropic to share Mythos cyber flaw findings with global finance watchdogglobal finance watchdog
Anthropic said it will share findings about a cybersecurity flaw in its Mythos model with a global finance watchdog and allow partners to pass related research to others.
The move, reported on May 18 and May 19, highlights growing concern that advanced AI systems can expose security risks across the financial sector.
- The Financial Stability Board is based in Basel, Switzerland.
- Anthropic was launched in 2021 by researchers who left OpenAI.
- Large language models can sometimes leak sensitive information even when they are not connected to the internet.
