A mock trial with the artificial intelligence ChatGPT acting as a judge was staged on Saturday by students at the University of Tokyo.
It was a test to see whether AI technologies are capable of making legal judgments and whether humans can accept judgements made by artificial intelligence in the future.
The mock trial scenario assumed a fictitious case in which a woman, who was harassed by her ex-boyfriend, was accused by prosecutors of conspiring with her current boyfriend to have her ex-lover killed.
The defendant's lawyer entered a plea of not guilty, saying the woman did not ask that her ex be murdered, and was not aware of the plan.
The prosecutors argued that based on her friend's testimony it was clear that she had planned the murder.
The AI judge found the defendant not guilty after questioning the defendant and witnesses using a synthesized voice.
The judge explained that even though the woman harbored anger towards the victim, it could not be proved beyond doubt that she had conspired with her boyfriend to kill the man.
The organizers say that they want people to think about whether they could accept judgements made by AI if they were involved in legal cases.