AI-generated content and accuracy

Much of what you read on Open Council Network, including meeting summaries, topic digests, and some transcript tooling, is produced or assisted by automated systems, including large language models and speech-to-text. This page explains what that means for accuracy and how to verify information.

What is AI-generated?

• Meeting summaries: narrative summaries on meeting pages are generated from transcripts, agendas, and related documents.
• Email and digest products: editorial-style text in newsletters or professional briefings may be assembled or drafted with AI assistance from the same underlying material.
• Transcripts: audio is converted to text using automated transcription; wording may not match the spoken record exactly.
• Search and chat features: where offered, responses are generated from retrieved source passages and may omit context.

Safeguards and limitations

We apply checks and human review in some workflows, but automated systems can still omit detail, merge topics incorrectly, or state things that are not supported by the source material (“hallucinations”). Summaries describe what appears in the available record; they are not legal advice, official minutes, or a council’s formal position.

What you should do

1. Read the source documents, agenda papers, and recordings linked on the meeting page.
2. Compare any claim that matters to you against the original transcript or minute.
3. Remember that a meeting may discuss earlier decisions or petitions without taking a new decision at that meeting.

Reporting errors

If you spot a factual error in a summary, email product, or transcript, we encourage you to tell us. Please include the meeting or page URL and a short note on what is wrong. Email community@opencouncil.network with the subject “Content accuracy report”.

We review every report and act on valid issues: we correct published summaries where appropriate, and source evidence on the site (documents, recordings) is generally left unchanged. Where a report shows a pattern or a gap in our safeguards, we also adjust prompts, checks, or other parts of the system so similar mistakes are less likely in future. Reporting errors helps improve the service for everyone.