LLM-powered robots are prone to discriminatory and dangerous behavior
New research indicates that robots operated by popular artificial intelligence systems are prone to enacting discriminatory, violent, and unlawful behaviors, making them unsafe for real-world use.