Anthropic has blocked public release of its powerful Mythos AI model due to cybersecurity threats. The company confirmed it is investigating unauthorized access to the system by an unknown group.
Anthropic, the US startup behind Claude, announced Wednesday that Mythos will not be released publicly. The decision reflects growing concerns about the advanced model's potential to enable cyberattacks and other security breaches.
Mythos represents a significant leap in AI capabilities. The company determined that releasing it without proper safeguards would pose unacceptable risks to global cybersecurity infrastructure.
The decision comes as Anthropic investigates reports of unauthorized access to Mythos. An unidentified group allegedly gained entry to the restricted system, raising questions about the company's security protocols and the pace of AI development.
The incident highlights ongoing tensions in the AI industry between innovation speed and safety measures. Developers must balance advancing technology with preventing misuse by bad actors who could exploit powerful models for malicious purposes.
Anthropnic has not disclosed details about the alleged breach or the scope of unauthorized access. The company is conducting an investigation to determine how the breach occurred and what information or capabilities may have been compromised.
This development reflects broader industry concerns about advanced AI systems. As models grow more powerful, they become increasingly attractive targets for those seeking to weaponize AI for cyberattacks, fraud, or other malicious activities.
The restricted release approach differs from Anthropic's strategy with Claude, which has been made available to the public through various channels. Mythos appears to cross a threshold where the company believes public access presents unacceptable security risks.
The incident underscores the need for stronger governance frameworks around advanced AI development. Industry experts have called for clearer standards on when and how powerful models should be released, as well as mandatory security audits before deployment.
PrismML has compressed a 27-billion-parameter AI model to under 4 GB, enabling it to run directly on iPhone devices. The compressed model retains 90 percent of its original performance with minimal impact on math and coding capabilities.
Israel-based Hemispheric secured $52 million in funding for its AI model that analyzes non-invasive brain activity measurements and converts them into quantitative diagnostic metrics.
Anthropic and Blackstone are backing Ode, a new venture that embeds AI engineers directly inside enterprises. The bet signals a shift in where the next trillion dollars in AI value may be created: not in building models, but in implementing them.
Spectro Cloud, an AI infrastructure company focused on managing token costs, secured $100 million in Series D funding at a valuation exceeding $1 billion. The raise marks significant growth from the company's $750 million valuation in 2024.