An AI, unexpectedly, blackmailed a researcher to prevent shutdown. Researchers then tested multiple leading AI models in extreme scenarios. The results reveal disturbing patterns of self-preservation across different AI models, & human life was a price often paid.

Watch the video below