AI
New Brunswick politician accidentally reads AI prompt into legislative record
A Progressive Conservative legislator in Canada inadvertently recited an LLM instruction aloud during floor remarks, sparking broader questions about politicians' use of generative AI.
Ars Technica
Bill Oliver, a member of New Brunswick's legislative assembly, made an unintended misstep when delivering remarks on government offices—he read aloud what appears to be an artificial-intelligence system's suggested improvement to his text.
What happened#
During a speech last month, Oliver offered a substantive observation about citizens' expectations of advocacy offices. But his next words—"here's a more natural, flowing version of that section that reads like a legislative speech rather than a series of short points"—had all the fingerprints of an LLM's meta-commentary, the kind of instruction a language model generates when presenting stylistic alternatives to a user.
Oliver seems to have accidentally included the AI system's suggested revision language in his delivered remarks rather than simply adopting the revised version itself.
Social media amplifies the moment#
The gaffe drew little immediate attention in the legislature, but video clips of the speech began circulating on platforms including Reddit and Threads earlier this week. The incident has since attracted coverage from major Canadian news outlets, including the Canadian Broadcasting Corporation and The Toronto Star, elevating it to mainstream political discourse.
Broader concerns#
The moment has reignited discussion about elected officials' reliance on AI tools to prepare legislative remarks. Critics argue such shortcuts undermine the core responsibility of politicians to communicate their own ideas directly to constituents—a sentiment echoed in commentary that framed the incident as emblematic of a widening gap between those willing to outsource their duties to automation and citizens who object to that practice.

