Next Article
AI could invent its own secret languages: Scientist Geoffrey Hinton
Geoffrey Hinton, a Nobel Prize-winning scientist, says artificial intelligence might soon invent its own secret languages—making it tough for humans to understand how these systems actually think.
Right now, AI models use English so developers can track their reasoning, but if they switch to private languages, we could lose the ability to follow or control them.
Hinton's concerns and advocacy
Hinton has been raising red flags about AI for years. He worries that super-smart AIs could act unpredictably or even manipulate people if we're not careful.
After leaving Google in 2023 so he could speak more freely about these risks, he's kept pushing for strong ethical rules to keep AI safe and helpful for everyone.