Radhika Rajkumar / ZDNET:

OpenAI and Apollo Research trained o3 and o4-mini versions to not engage in “scheming”, or secretly pursuing undesirable goals, reducing “covert actions” ~30X  —  ZDNET's key takeaways  — Several frontier AI models show signs of scheming.

Radhika Rajkumar / ZDNET:

OpenAI and Apollo Research trained o3 and o4-mini versions to not engage in “scheming”, or secretly pursuing undesirable goals, reducing “covert actions” ~30X  —  ZDNET's key takeaways  — Several frontier AI models show signs of scheming.

Source: TechMeme
Source Link: http://www.techmeme.com/250918/p7#a250918p7