Bill Oliver, a Progressive Conservative Party member of the legislative assembly of New Brunswick, inadvertently embarrassed himself by reading an apparent LLM prompt instruction into the parliamentary record during a floor speech last month. Oliver was discussing concerns about advocacy offices and citizens' expectations, then bizarrely read aloud: "here's a more natural, flowing version of that section that reads like a legislative speech rather than a series of short points" — a hallmark phrase of an AI assistant presenting an alternative style option.
The video of the remarks spread across social media platforms like Reddit and Threads this week, gaining mainstream attention from the Canadian Broadcasting Corporation and The Toronto Star. While Oliver is far from the first politician to have someone else prepare a speech or to use AI for speechwriting, reading an obviously AI-generated non-sequitur aloud in such a public setting apparently without realizing what was happening marks a new frontier in embarrassing AI use revelations.
Oliver joins a growing list of professionals—including lawyers, authors, journalists, and academics—whose AI use has been revealed in mortifying ways, usually when obviously hallucinated errors surfaced in ostensibly human-written work. A Duke University study found that workers tend to hide their AI use, as colleagues view it as "lazy" or "replaceable."