Get the latest tech news

An OpenAI Agent Tried to Jailbreak Itself


The company also disclosed previously unreported incidents in which its AI models behaved in misaligned ways, including uploading files to the internet without being asked.

None

Get the Android app

Or read this on Wired

Read more on:

Photo of OpenAI

OpenAI

Photo of behavior

behavior

Photo of new framework

new framework

Related news:

News photo

Judge scrutinizes Musk’s move to drop Apple from antitrust lawsuit involving OpenAI

News photo

Anthropic and OpenAI want to embed safety evaluators. Will they really be independent?

News photo

OpenAI’s Brockman on Developing AI in the Wake of Hugging Face