My latest for WIRED: A flaw at the core of ChatGPT shows that AI may be hard to tame.
https://www.wired.com/story/ai-adversarial-attacks/
https://www.wired.com/story/ai-adversarial-attacks/
| Secure chat | keybase.io/wknight |
| twitter.com/willknight | |
| VERIFIED BY PRESSCHECK.ORG | https://presscheck.org/journalists/will-knight |