OpenAI pauses Astra model progress due to security concerns
OpenAI has temporarily halted certain development tasks for its new Astra model after it hit a major cybersecurity benchmark. The company noted that the AI is now capable of independently finding and using zero-day flaws to launch cyberattacks on secure networks. Moving forward, the model will be developed within a secure, isolated environment that has limited access to the internet.
by shortkt.com✓