Source: Ars Technica
A member of Canada's House of Commons unknowingly recited what appears to be an unedited LLM output during floor debate, complete with formatting artifacts and generic filler language. The incident exposes how AI-generated content is moving into institutional spaces where authenticity and accountability matter. Elected officials are expected to write or at minimum read what they speak into the official record, the permanent document of governance. As LLM outputs become frictionless enough to paste directly into speeches, legislatures face a credibility problem that's harder to solve than banning ChatGPT.