Anthropic suspends access to new AI model by order of the US; find out more

Anthropic has suspended access to its new artificial intelligence models, Fable 5 and Mythos 5, for foreign users. The measure would have been taken by order of the United States government. The decision applies to any citizen of another country, even if they are within American territory, and also affects foreign employees of the company itself.

According to the company, the order was issued based on national security issues and does not specify which risks motivated the decision.

In a statement, Anthropic said it understands that authorities believe they have identified a way to bypass Fable 5's security protections, but said the technique presented involves vulnerabilities considered limited and that similar capabilities can be found in other publicly available models.

The company also stated that it analyzed a demonstration of the supposed technique and found only minor flaws, with no specific benefits related to the model. No test has so far found a universal method capable of breaking the system's security barriers.

Anthropic argued that Fable 5 was developed with advanced protection mechanisms to reduce the misuse of technology in activities related to digital attacks. According to the company, the model underwent thousands of hours of security tests before launch, including evaluations with government agencies and private organizations.

The other Claude models continue to operate normally. The US government made no further statements.

What is Claude
Founded in 2021 by former OpenAI employees, Anthropic is one of the main companies in the race for generative artificial intelligence. The company is responsible for the Claude chatbot, which offers similar features to ChatGPT.

Despite being a recent company, Anthropic has received billion-dollar investments from technology giants, such as Google and Amazon, and has gained prominence for defending a security-oriented approach in the development of AI systems.

The company states that its mission is to create more reliable models with greater control over possible misuse.

Young people are among the most suspicious of AI, says global survey

Source: CNN