%20(1).jpg){kind=small}
The debate over Anthropic AI security risks intensified after Amazon CEO Andy Jassy raised concerns with senior Trump administration officials. According to Reuters, Jassy was among several technology leaders who discussed potential security issues linked to Anthropic’s most advanced artificial intelligence models.
The development highlights growing attention on Anthropic AI security risks and government oversight of advanced AI systems.
Anthropic Suspends Advanced AI Models
Anthropic took the unusual step of disabling access to its latest models worldwide. The decision followed national security orders issued by President Donald Trump’s administration.
The San Francisco-based startup had recently introduced a public version of its advanced model. The model, called Fable, was released with cybersecurity protections. Before that release, Anthropic had delayed wider deployment of its Mythos model due to concerns about hacking capabilities.
The public release lasted only a few days before the company withdrew access.
According to Anthropic, U.S. officials identified a possible method for bypassing model safeguards. The company described the issue as a jailbreak technique.
Government Orders Global Restrictions
Anthropic stated that the administration ordered the company to prevent all foreign nationals from accessing Fable 5 and Mythos 5. The restriction applied both inside and outside the United States.
As a result, Anthropic disabled access globally.
Key points from the order include:
- Restrictions on foreign nationals worldwide.
- Suspension of Fable 5 and Mythos 5 access.
- Export control measures issued by the U.S. government.
- Concerns about potential safeguard bypass methods.
The company said the identified bypass exposed only minor security flaws. It also noted that similar issues can be found by other publicly available AI models.
Amazon and Government Discussions
Amazon did not confirm whether it specifically discussed Anthropic’s models with officials.
A company spokesperson stated that governments often seek advice from Amazon because it serves many public and private sector customers. However, the spokesperson declined to share details of any discussions.
Reports from The Information indicated that Jassy expressed concerns regarding Anthropic AI security risks during conversations with government officials.
Export Controls Spark Debate
Anthropic said the restrictions were issued through export control measures. The U.S. Commerce Department's Bureau of Industry and Security oversees such controls.
White House adviser David Sacks stated that officials acted reluctantly. He claimed Anthropic CEO Dario Amodei declined to fix the jailbreak issue or remove the model from service.
Sacks said the administration hopes Anthropic addresses the problem so restrictions can be lifted and Fable can return to general release.
Some experts questioned the policy. They argued that the controls affect allied countries as well as potential adversaries. Jimmy Goodrich of the University of California called the decision surprising and said it was not thoroughly planned.
The dispute over Anthropic AI security risks continues as policymakers, technology companies, and researchers assess the balance between innovation and national security.
