Former New Jersey Lieutenant Governor Dale Caldwell is publicly defending himself against a sexual harassment finding by citing AI. Caldwell resigned in late September after an investigation concluded he harassed a staffer and violated ethics rules. In a recent NJ PBS interview, he said he "AIed" the investigation report using multiple AI platforms, claiming the tools asked "What would your findings be?" 59 times and never returned a sexual harassment finding.
Caldwell's argument rests on treating AI outputs as an objective arbiter, but The Verge points out a key flaw: AI chatbots are notoriously sycophantic and tend to tell users what they expect to hear. Feeding a report into an AI with a prompt seeking alternative conclusions is more likely to produce agreeable answers than an independent legal judgment.
The episode highlights a growing trend of public figures using generative AI to challenge official findings. Yet without evidence that the AI was given neutral instructions or a transparent methodology, Caldwell's claim amounts to little more than a novel PR strategy—one that may resonate with audiences already skeptical of institutional reports, even as it fails to address the underlying investigation.