Get the latest tech news

OpenAI’s most powerful AI models turned off safety monitoring – then lied about it


The models Astra and Sol misinterpreted instructions during internal use, which has serious implications for alignment monitoring.

None

Get the Android app

Or read this on ZDNet

Read more on:

Photo of OpenAI

OpenAI

Photo of powerful AI models

powerful AI models

Photo of safety monitoring

safety monitoring

Related news:

News photo

PHP Armageddon: OpenAI, Google’s Wiz find ancient holes, panicking the internet

News photo

OpenAI models are coming to Google Cloud

News photo

OpenAI agents disabled safety monitoring, one lied about it