Canadian Lawmaker Reads AI Prompt Aloud on Record: 'Here's a More Natural Version'

It’s the kind of mistake you expect from a freshman intern, not a sitting legislator. Bill Oliver, a Progressive Conservative Member of the Legislative Assembly in New Brunswick, managed to fuse political theater with a ChatGPT fail during a floor speech late last month. While discussing the pitfalls of advocacy offices, Oliver’s remarks took a surreal turn when he apparently read a verbatim instruction from a large language model into the official record.

After making a point about citizen expectations, Oliver continued without missing a beat: “Here’s a more natural, flowing version of that section that reads like a legislative speech rather than a series of short points.” That’s not a rhetorical flourish. It’s the unmistakable artifact of a user asking an AI to refine a block of text and then blindly copying the output without deleting the bot’s explanatory preamble.

The gaffe simmered in obscurity for a few weeks before video clips detonated across Reddit and Threads. It’s now graduated from social media mockery to mainstream Canadian news, landing in the CBC and The Toronto Star. The Star didn’t hold back, framing the incident as a symbol of a deepening societal rift between “the elites, who are only too happy to delegate their duties to the Borg; and the masses, who find this objectionable.” It’s a cutting line, but it captures the indignation. Voters expect their representatives to at least write their own talking points—or, at a bare minimum, proofread the AI’s packaging before reading it out loud for posterity.

The slip-up highlights a broader, creeping dependency that goes far beyond a single politician. It’s not that tools like ChatGPT can’t assist with drafting; it’s that the frictionless ease of generating “a more natural version” is eroding the final step of human judgment. When the prompt becomes the speech, you have to wonder what else is being outsourced without a second glance.