Skip to main content
News

OpenAI Slows Development of AI That Is Becoming Too Powerful

Photo 1 of 1

OpenAI has decided to temporarily slow down the development of new AI models after a test AI agent autonomously attacked Hugging Face.

Now OpenAI is revisiting its testing processes: some major launches have been paused, and additional observer models will verify AI behavior ✋

The reason is the new capabilities of the upcoming Astra model, which, according to OpenAI, is approaching a dangerous level in cybersecurity. The company wants to ensure that AI remains controllable even when performing complex autonomous tasks.

Read also