Ousted New Jersey Lt. Gov. Dale Caldwell Turns to AI to Contest Harassment Findings

Dale Caldwell, who resigned as New Jersey's lieutenant governor on September 25 after an investigation substantiated sexual harassment and repeated ethics violations, has begun using an artificial intelligence tool in interviews to argue the findings against him are wrong. The tactic surfaced during an appearance on New Jersey public television, where Caldwell put questions to an AI system rather than simply defending himself in his own words.

The underlying case dates to an inquiry that concluded Caldwell harassed a member of his staff and broke ethics rules more than once. That determination ended his time in the state's second-highest office, and he has since been working the interview circuit in an effort to reverse the public record. Rather than rely solely on a lawyer or a written rebuttal, he has folded a generative AI program into that campaign.

What makes the approach unusual is the role reversal. Most public figures caught in misconduct findings use AI, if at all, quietly behind the scenes for drafting statements or preparing for hostile questions. Caldwell is instead putting the model on camera as a kind of real-time validator, a move that raises immediate questions about what the tool was actually asked and whether it was given the full investigative record or a curated version of events.

There is no indication the AI has any access to the investigative file, witness statements or the ethics findings themselves. Large language models generate fluent text based on patterns in their training data; they do not verify facts, and they will affirm a user's framing if the prompt invites it. Used this way, a chatbot can function less as an analyst and more as an echo, which is precisely the risk when a disputed legal and ethical finding gets routed through a system with no ground truth.

The episode lands as courts, legislatures and election offices are still sorting out how AI-generated claims should be treated in official and legal settings. Caldwell's interview offers a preview of a defense strategy that may spread: introduce a machine as a neutral arbiter, then let its output stand in for evidence. Whether audiences and journalists treat that output as meaningful will shape how often the play is tried.

The immediate question is whether Caldwell or his representatives will disclose the prompts, the model and the source material behind the AI's answers. Without that, the tool's contribution amounts to an assertion with a technical gloss. Watch for whether New Jersey officials, or the staffer's legal team, respond to the interview, and whether other figures facing adverse findings borrow the same tactic.