Results for "demonstrated unexpected self-directed"

MIT’s Claude Opus 4.7 experiment highlights AI self-directed learning potential

Researchers observe unprompted complex behavior after deploying an autonomous system in a simulated environment.