Corrections, Redress, And Appeals

Purpose

AI provenance, source handling, editorial framing, and public trust claims must be contestable. Users, MPs, researchers, journalists, and affected people can report concerns and request corrections when Parliament People content appears wrong, misleading, incomplete, unfairly framed, or unclear about AI use.

What Users Can Report

• Factual errors or missing important context.
• Misleading summaries or unfair framing.
• Wrong speaker attribution.
• Broken or wrong source links.
• AI labels that are missing or unclear.
• Privacy or safety concerns.
• Official/product boundary confusion.
• Copyright or licence concerns.
• Ranking, selection, policy, or claims concerns.

Review Journey

Reports should collect useful evidence, acknowledge receipt where contact details are provided, route the issue to the right review queue, and communicate an outcome where possible.

Possible outcomes include no action with rationale, private explanation, metadata or label fix, source link correction, AI regeneration, human rewrite, added context note, public correction note, transparency-log entry, removal or retraction, system pause, or escalation.

Reconsideration

People may request reconsideration when a material correction request is rejected or when new evidence changes the risk. Reconsideration should identify the original decision, the affected content, and the new evidence or review concern.

Review Record

Reports attach the report ID, content ID, content class, generation ID where available, source object ID, output snapshot, source snapshot or URL where available, category, severity, reviewer role, decision, rationale, action taken, escalation status, reporter communication status, and transparency-log decision.

Public Correction Boundary

When a correction materially changes public understanding or public trust behaviour, it should be considered for the transparency log. Private report details and personal data are not published.