Canadian legislator accidentally reads LLM instruction in floor speech

Bill Oliver, a Progressive Conservative Party member of New Brunswick's legislative assembly, delivered a floor speech last month that included substantive remarks on the dangers of advocacy offices. "One of the dangers associated with creating advocacy offices," he stated, "is that citizens often develop expectations that exceed the powers actually granted to those offices." However, what followed was not a continuation of his prepared remarks but an apparent LLM instruction that Oliver 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."

The phrasing is characteristic of how large language models present alternative text options to users in response to a prompt, suggesting that Oliver's speech had been drafted or edited using an AI tool and that he failed to remove the instructional text before reading from the script. The remarks initially attracted little notice when delivered, but video of the speech surfaced on social media platforms including Reddit and Threads earlier this week, sparking wider circulation.

Canadian mainstream media outlets, including the Canadian Broadcasting Corporation and The Toronto Star, have now picked up the story. The Toronto Star characterized the incident as emblematic of "a growing divide in our society: between the elites, who are only too happy to delegate their duties to the Borg; and the masses, who find this objectionable." The coverage frames Oliver's apparent reliance on AI-generated text as a symptom of public officials outsourcing legislative speech without sufficient care, raising broader questions about accountability and authenticity in government communication.

The incident reflects a broader pattern of public figures using AI tools without adequate quality control. Oliver's speech included a substantive statement about the risks of advocacy offices — a legitimate legislative topic — but the accidental inclusion of the LLM instruction reveals that the content was drafted or refined using an AI text generator. The fact that the speech went largely unnoticed when delivered but gained traction only after video spread on social media suggests that the problem may be widespread but often invisible until amplified by public scrutiny.

The Toronto Star's framing of the incident as evidence of a "divide" between elites and the general public speaks to a deeper concern: officials who outsource their communications to AI may be seen as abrogating their duty to engage thoughtfully with constituents. For legislators especially, the use of AI-generated language without careful review risks undermining public confidence in the authenticity and care of official statements.

No comments yet. Be the first to share your thoughts!