Artificial intelligence start-up company Anthropic has announced a new A.I. model but is limiting its rollout over fears that it could enable hackers to exploit software vulnerabilities.
The model, called “Claude Mythos Preview,” excels at identifying weaknesses and security flaws within software products.
As a result, Anthropic is limiting access to Claude Mythos in an effort to prevent bad actors from exploiting those capabilities and launching cyberattacks around the world.
Rather, Anthropic is making Claude Mythos available to a select group of U.S. companies as part of a new cybersecurity initiative called “Project Glasswing.”
The companies Anthropic is granting Claude Mythos access to are Apple (AAPL), Alphabet (GOOGL), Microsoft (MSFT), Nvidia (NVDA), JPMorgan (JPM), and CrowdStrike (CRWD).
In all, about 40 companies are being given access to the powerful new A.I. model.
News of the Claude Mythos A.I. application sent stocks of cybersecurity and software companies lower.
Anthropic was founded in 2021 by a group of researchers who left OpenAI over concerns about its direction and attitude towards safety.
The company has spent years constructing its reputation as a firm that is more dedicated to responsible A.I. deployment.
Anthropic says it hopes leading U.S. companies will use Claude Mythos to find bugs and vulnerabilities in their software and fix them.
Management also said they are in active discussions with the U.S. government about Claude Mythos, the potential for future cyberattacks, and ways to safeguard critical infrastructure.
Anthropic is a privately held company and its stock does not trade on a public exchange. However, the A.I. start-up is expected to go public later this year.
Related Stories
