❗️Microsoft’s CEO says AI models should be treated as already COMPROMISED.

“We must assume a model is compromised and contain it from the start,” Satya Nadella says.

He wants an authorized person to be able to “pause or shut down a model mid-task,” like an “emergency brake.”

He’s the latest tech leader to speak out on AI safety, after Elon Musk and Sam Altman backed Dario Amodei’s call to slow down AI.

@aitopics