Most humans learn the skill of deceiving other humans. So can AI models learn the same? Yes, the answer seems — and terrifyingly, they’re exceptionally good at it. A recent study co-authored by researchers at Anthropic, the well-funded AI startup, investigated whether models can be trained to deceive, like injecting exploits into otherwise secure computer […]
© 2023 TechCrunch. All rights reserved. For personal use only.
via https://www.aiupnow.com
Kyle Wiggers, Khareem Sudlow
