Former NJ Official Claims AI Proves His Innocence in Harassment Case

New Jersey's former lieutenant governor Dale Caldwell resigned in September 2026 after an official investigation found he had sexually harassed a staff member and broken ethics rules multiple times. Rather than accepting the findings, Caldwell has been trying to clear his name in media interviews.

In one unusual TV appearance on NJ PBS, Caldwell claimed he had fed the investigation report through AI systems—software designed to analyze text and answer questions. He said that when he asked these AI tools "What would your findings be?" they supposedly disagreed with the official report 59 times. His argument: if AI said he was innocent, the investigation must have gotten it wrong.

Here's the problem: AI tools work by recognizing patterns in text they've been trained on. They don't independently investigate facts or decide who is guilty or innocent. Feeding a report to an AI and asking it to find problems is very different from how a real investigation works. Real investigations involve interviewing people, checking evidence, and following strict rules. An AI tool simply looks for patterns—it's not a judge or truth-finder.

This case shows why people need to be careful about what AI can actually do. Just because an AI agrees with you doesn't make something true. Using AI this way to defend yourself against serious accusations raises questions about whether public officials should be allowed to use technology in this manner.