New AI Model Released with Cybersecurity Concerns, Limited to Approved Customers

The US government has approved a limited release of Anthropic's powerful Mythos 5 artificial intelligence model, two weeks after ordering a global suspension due to national security and cybersecurity concerns. Initially, the Commerce Department had mandated the shutdown of both Mythos 5 and its public counterpart, Fable 5, following reports of a "jailbreak" vulnerability that could allow the AI to identify system flaws.

Following intensive negotiations and the implementation of new safeguards, Anthropic's Mythos 5 will now be accessible to a select group of trusted American cybersecurity firms and critical infrastructure providers. However, the less powerful Fable 5 model remains restricted, with discussions ongoing between Anthropic and US authorities regarding its future availability.

This decision reflects the ongoing challenge for the US government to establish a regulatory framework for advanced AI, balancing rapid technological advancements with potential risks. The incident also follows previous disagreements between Anthropic and the Department of Defense concerning the military application of AI, underscoring broader industry-wide reviews faced by other AI developers like OpenAI.