Current evidence points to modest effects of artificial-intelligence tools on jobs, highlighting how bad data, not sweeping automation, is driving much of today’s alarm.

Large language models appear aligned, yet harmful pretraining knowledge persists as latent patterns. Here, the authors prove current alignment creates only local safety regions, leaving global ...

Mar. 26, 2026 Deepfake X-rays created by AI are now convincing enough to fool both doctors and AI models. In tests, radiologists had limited success identifying fake images, especially when they ...