At the Chicago School of Law, 61 US federal judges were replaced by GPT-5 in a test on traffic accidents with different conditions. The result is impressive: AI applied the correct norm 100% of the time, while humans erred in about half the cases.

GPT-5 completely ignored emotions and subjectivity, acting strictly by the letter of the law.

Judges were more likely to explore "gray areas" to achieve a fairer outcome, while AI never deviated. An ideal formal judge — is this progress or a loss of human judgment?

By the way, in our Neiron bot, an even more advanced model 5.2 is available in Premium plans ⚡️

AI Decided Cases Better Than Real Judges