Horizon Reports is committed to accuracy in all our AI-generated journalism. When errors occur — and they will — we correct them quickly, transparently, and with full accountability.
We believe corrections are not a failure — they are proof of accountability.
While our AI pipeline includes multiple quality gates to catch errors before publication, no system is perfect. When readers identify inaccuracies, we investigate, verify, and correct them according to the standards outlined on this page.
We classify corrections into three categories based on severity and impact.
Our AI pipeline is designed to catch most errors before publication. When issues slip through, here's how we respond.
AI quality gates catch 80-85% of potential errors before articles go live.
Reader-reported errors are reviewed by human oversight within 24-48 hours.
If you spot an error in one of our articles, please let us know. We review every report.
corrections@horizonreports.org
Fastest for urgent corrections
Use our contact page
Include article URL and details
@hr_reports (X/Twitter)
Public corrections via DM
We aim to respond to all error reports within the following timeframes:
Fixed silently, no formal notice
Editor's note added to article
Article removed or fully rewritten
Note: These are target response times. Actual correction time may vary based on complexity of verification required. We will acknowledge receipt of your report within 24 hours.
We are committed to maintaining a public record of significant corrections and retractions.
Our public corrections log is currently under development. Once live, all significant corrections and retractions will be listed here with full details, including the original error, correction made, and date of update.