Researchers hacked several robots infused with large language models, getting them to behave dangerously—and pointing to a bigger problem ahead.

---

Link :
https://www.wired.com/story/researchers-llm-ai-robot-violence/