AI developer Anthropic has announced its latest Claude AI model, dubbed Claude Mythos, is too powerful and potentially dangerous for public use. The company has decided not to release the software to the general public.
Claude Mythos is described as a frontier AI model with capabilities substantially beyond previous versions in areas like software engineering, reasoning, and research assistance. Notably, it demonstrates advanced cybersecurity skills applicable to both defending and exploiting software vulnerabilities.
Anthropic stated that due to its significant increase in capabilities, Claude Mythos will be integrated into a defensive cybersecurity program with a select group of partners focused on maintaining critical software infrastructure. Experts emphasize the real-world national security implications of such advanced AI technologies.
The decision to withhold public access is viewed by some as a responsible measure to fix potential flaws before widespread deployment, akin to a pharmaceutical company holding back a new drug. This cautious approach aims to prevent the technology from falling into the wrong hands and being used to harm society.
Concerns remain about the rapid growth of AI and its potential impact on cybersecurity threats, such as ransomware attacks, which are becoming cheaper, faster, and harder to detect. The lack of clear governance and regulatory frameworks for advanced AI development is a growing issue.