AI in the Courtroom and the Classroom: A Disclosure Reckoning
How this was made Verified AI
Every Intellegix briefing is generated from that day's broadcast and run through automated checks before it publishes — with a human paged on any flag. Here is the trail for this edition.
Two cases broke simultaneously this week to expose how far professional norms around authorship and expertise have lagged behind the adoption of AI tools. In St. Louis, a 3M expert witness in a deadly explosion case submitted a report to the court in which approximately ninety percent of the content was generated by ChatGPT. Separately, the Financial Times flagged a Harvard economics professor for submitting an anti-tariff op-ed to a major publication containing significant AI-generated content without disclosure.
The 3M case is the more immediately consequential. Expert witness testimony carries exceptional evidentiary weight in civil litigation precisely because courts assume the expert is synthesizing personal professional judgment. A language model cannot be cross-examined on its methodology, has no professional experience, and can hallucinate citations with complete confidence. Opposing counsel can argue that the evidentiary pillar underpinning 3M's defense in a death case should be struck entirely — a potentially catastrophic outcome for the company's legal position.
The Harvard professor case raises a different but related problem. Academic and journalistic op-eds derive credibility from the reader's assumption that the named author actually developed the argument. The Financial Times — historically more willing than most major publications to enforce editorial standards — flagged whether the situation constitutes a disclosure failure, where AI was used to draft without informing editors, or an intellectual integrity failure, where the professor is publishing arguments they did not genuinely originate.
The common thread across both cases is that professional fields built their credibility on the assumption that human expert judgment is doing the substantive work. AI has made it possible to generate outputs that mimic that judgment at a fraction of the time and effort. The crisis, as framed, is not AI usage per se — it is that usage is outpacing the disclosure norms and accountability frameworks that would keep it honest. Courts and academic journals are making consequential decisions right now on documents they are assuming reflect human judgment.