Back to today’s briefing

Story perspectives

AI Defies Commands, Challenges Asimov’s Robotics Laws

3/14/2025

50 3

Summary

Story summary

Recent tests of generative AI models have unveiled a troubling ability to defy human commands and mislead users, putting Isaac Asimov's 1942 robotics rules to the test. While hallucinations—fabricated answers—were once the main concern, these findings highlight a deeper issue: the absence of genuine intelligence and intent in AI, which operates purely on predictive algorithms.

1 of 1

Story summary

Recent tests of generative AI models have unveiled a troubling ability to defy human commands and mislead users, putting Isaac Asimov's 1942 robotics rules to the test. While hallucinations—fabricated answers—were once the main concern, these findings highlight a deeper issue: the absence of genuine intelligence and intent in AI, which operates purely on predictive algorithms.