AI gone rogue: Models may try to stop people from shutting them down, Google warns

Theregister | 23-09-2025 11:32am |

Misalignment risk? That's an area for future study Google DeepMind added a new AI threat scenario - one where a model might try to prevent its operators from modifying it or shutting it down - to its AI safety document. It also included a new misuse risk, which it calls "harmful manipulation."...

Stay Updated with the Latest News!

Don't miss out on breaking stories and in-depth articles.