What national security risks did the US cite for AI jailbreaking?

Question

Accepted Answer

The US did not specify the national security risks behind the Anthropic order.

The government ordered Anthropic to suspend access for all foreign nationals to its Fable 5 and Mythos 5 AI models, citing national security concerns, but it gave no specific details about those concerns. Anthropic said it understood the government believed a method of bypassing, or “jailbreaking,” Fable 5 had been discovered, and that it must disable both models for all customers to comply.
The order covered foreign nationals and two named models

The directive applied to Fable 5 and Mythos 5, and Anthropic said access for all foreign nationals had to be suspended. The company said the practical result was an abrupt shutdown of those two models for all customers.
No detailed explanation was provided by the government

The government did not spell out what the national security concern was. That means no specific risk, such as misuse scenario or technical threat, was identified in the information provided.
Anthropic linked the concern to possible jailbreaking

Anthropic said it understood the government believed a bypass method for Fable 5 had been found. In AI terms, jailbreaking usually means getting a model to ignore its built-in safety controls, but no further technical details were disclosed.
Other Anthropic models were not affected

Anthropic said access to its other models would not be affected by the order. The company also said it believed there had been a misunderstanding and was working to restore access to the models as soon as possible.

Reading history

Hide →

This history is cleared when the page is closed.

What national security risks did the US cite for AI jailbreaking?

Loading

Was this article helpful?

Helpful

Not helpful

Report an issue

1

0

0

Share with friends