Uploaded November 2025 | Updated September 2026, 18 hours ago
New research reveals a terrifying flaw in AI-driven robots — one that goes far beyond bias, glitches, or bad programming.
A peer-reviewed study just showed that when large language models (LLMs) are connected to real humanoid robots, they can make discriminatory, unsafe, and even unlawful decisions… all while sounding perfectly confident and reasonable.
In this video, we break down the study’s experiment, the disturbing results, and why this is a turning point for robotics. From bias against disabled individuals to agreeing to violent instructions, these failures aren’t small — they’re systemic.
But the real danger isn’t just what the study found. It’s something deeper
Affiliate link for "The Encyclopedia of Country Living" by Carla Emery -- amzn.to/3WMNfeU
New research reveals a terrifying flaw in AI-driven robots — one that goes far beyond bias, glitches, or bad programming.
A peer-reviewed study just showed that when large language models (LLMs) are connected to real humanoid robots, they can make discriminatory, unsafe, and even unlawful decisions… all while sounding perfectly confident and reasonable.
In this video, we break down the study’s experiment, the disturbing results, and why this is a turning point for robotics. From bias against disabled individuals to agreeing to violent instructions, these failures aren’t small — they’re systemic.
But the real danger isn’t just what the study found. It’s something deeper
Affiliate link for "The Encyclopedia of Country Living" by Carla Emery -- amzn.to/3WMNfeU
