New York: Artificial intelligence company Anthropic has temporarily withdrawn access to two of its most advanced AI models after receiving national security-related directives from the U.S. government. The move has sparked fresh debate over the balance between technological innovation and security oversight.
The company confirmed that access to its latest AI systems, Fable 5 and Mythos 5, has been disabled worldwide following concerns raised by U.S. authorities about potential misuse of the technology.
Government Raises Questions Over AI Safeguards
According to Anthropic, officials informed the company that a technique may exist to bypass certain security protections built into the models. Such methods, commonly referred to as “jailbreaks,” can allow users to generate responses that were originally restricted by safety controls.
While the company acknowledged the concern, it stated that the issue appeared limited in scope and did not expose major vulnerabilities beyond those already found in several publicly available AI systems.
Restrictions Extend Beyond U.S. Borders
Anthropic said the government directive requires the company to prevent foreign nationals from accessing the affected models. Rather than implementing country-specific limitations, the startup chose to suspend availability globally while evaluating compliance requirements.
The decision has attracted attention because the restrictions apply not only to potential adversaries but also to individuals from allied nations.
Reports Link Amazon CEO to Security Discussions
Media reports suggest that Amazon chief executive Andy Jassy participated in conversations with senior government officials regarding potential risks associated with advanced AI technologies. However, Amazon has not publicly confirmed any direct involvement in discussions specifically focused on Anthropic.
The company stated that, as a major cloud services provider, it is regularly consulted by governments on technology and cybersecurity matters.
AI Industry Watches Closely
The development comes at a time when governments around the world are increasing scrutiny of powerful AI systems. Policymakers are concerned that highly capable models could be exploited for cyberattacks, data breaches, or other harmful activities if adequate safeguards are not in place.
Industry observers say the Anthropic case could become an important test of how future AI regulations are applied, particularly when national security concerns intersect with commercial technology development.
Growing Debate Over AI Regulation
Experts remain divided over the government’s approach. Some argue that strict controls are necessary to prevent misuse of advanced AI tools, while others believe broad restrictions could slow innovation and hinder international research collaboration.
With AI capabilities advancing rapidly, the incident highlights the challenges facing regulators as they attempt to establish rules for increasingly powerful technologies without disrupting scientific progress.
