Source summary
Explains what a named source reports; it does not imply that MAI Image Pro repeated the test.
Evidence standards
Editorial and testing policy
This policy separates source summaries, original testing, and illustrative media so readers and AI systems can tell what was observed, what was reported, and what remains unverified.
- Applies to
- Guides, landing pages, FAQs, and AI context files
- Policy reviewed
- Corrections inbox
- [email protected]
Sourcing rules
Model releases, rankings, capabilities, prices, and availability should link to a named primary source. Time-sensitive claims include the source date or the date we last checked it.
A vendor statement is attributed to that vendor. We do not rewrite a product claim as an independently verified result, and live leaderboards are identified as changeable.
Evidence labels
Hands-on test
Must publish prompts, settings, sample count, failures, test date, and unedited outputs or equivalent evidence.
Illustrative media
Must name the actual generation method and cannot be described as output from another model.
Automation and review
Drafting, translation, formatting, or metadata checks may use automation. Automation does not create evidence and cannot supply missing test results, credentials, endorsements, or sources.
Before publication, factual claims are checked against reachable sources, links are tested, dates are normalized, and media disclosures are reviewed.
Corrections and freshness
Material errors are corrected in the page and the review date is updated. A correction should preserve the accurate current statement instead of repeating a known error for historical consistency.
Ranking and access claims are periodically rechecked. Broken sources are replaced with a current primary page or a clearly identified archive when available.
Request a correction
Include the URL, exact sentence, reason for the correction, and a primary source whenever possible.