Paul Stenhouse: Should we be worried about AI breaking containment?
The AI is now breaking out of its containment. Should we be afraid?
Boy, it's almost like Hollywood predicted this! Our worst fears about AI have come true, with OpenAI saying its new ChatGPT model managed to connect to the internet and hack into another AI company's servers. It was meant to be contained, it was not meant be connected to the internet, but it found a way through. These new models are so good, they're finding vulnerabilities and exploiting them all on their own to achieve a goal.
JPMorgan Chase's CEO Jamie Diamond was one of the people briefed into Anthropic's powerful model…
The skinny
The skinny isn't ready yet — notes appear once the transcript is processed.