AI Verification

URX runs automated verification (model-assisted) on uploaded images. Signals include manipulation risk, date plausibility, alignment with declared distance/time, and internal consistency with known tracker layouts.

Possible outcomes:

• Verified at automation—still subject to later human review if policy triggers (see Manual review).
• Pending or manual review when the model is unsure.
• Declined when checks fail clearly—including sessions outside the three-calendar-day freshness rule in Submission requirements, or evidence that does not match an eligible activity.

Automation can err in both directions. It is a first pass, not the last word on its own—manual review and policy remain part of the same verification chain.

See also: Manual review; Submission requirements; Enforcement pipeline.