The company Anthropic stops 2 new AI models. The names are Fable 5 and Mythos 5. The US government orders this. It worries about the security of the country.
Researchers at Amazon test the AI. They trick the security locks. This is a Jailbreak. The AI tells secret information. It gives tips for bugs in computer programs.
Anthropic does not like the ban. The company says: There is no big error. Experts for Computer-Sicherheit are also angry. They need the AI for their work. The AI helps to find bugs.
There is also trouble with the military. The Pentagon wants to use the AI. But Anthropic does not want to build weapons. Now other countries cannot use the AI.
A report says: China maybe had access to Mythos. This is very dangerous for the USA. China could copy the technology. But this is not yet proven.