When AI Lands in the Contract: Nurses Bargaining Over AI

Part of Nurse.org’s Nursing AI Watch, our ongoing investigation where we track how artificial intelligence is reshaping bedside nursing. Here, we look at something few others are tracking: the first nurse union contracts to write explicit rules for AI at the bedside.

As AI becomes a standardized part of care, nurses are writing AI rules into their union contracts. In some areas, they are even winning safeguards on how the technology gets used. The pattern started in two states, and it’s likely to be replicated in others. 

However, the contract integration didn’t happen because employers volunteered it. In most cases, it took a strike, a credible strike threat, or years of organizing to get AI onto the bargaining table at all. The seven systems where nurses have now ratified AI contract language represent a small fraction of the health systems deploying AI tools. But they’ve established a template, and templates travel.

Here’s what the language says and who has it.

Want to see more Nurse.org articles in your Google results? Add us as a preferred source.

How Did AI First Land in a Nurse Contract?

Over two rounds of bargaining — one in New York, one in California — nurses turned vague worries about AI into binding contract language. Across those two waves, we’ve identified 7 health systems where nurses now have ratified AI or technology protections in their agreements.

That puts nurses among a small but growing set of workers setting hard rules on how AI gets used at work. Nursing isn’t first to this fight — Hollywood writers and actors negotiated the highest-profile AI protections in 2023 and 2024, with the Writers Guild of America barring AI from writing scripts and SAG-AFTRA securing consent rules for AI voice and likeness replicas. Journalists, tech workers, and teachers have since followed.

But healthcare is one of the fields where the stakes are clinical, not just creative, which makes the nurse contracts an early and closely watched test case. Most workers still have no say over workplace AI. These nurses negotiated one.

The first wave of nurse movement to address AI in their contracts came out of New York. The New York State Nurses Association (NYSNA) bargained AI safeguards into contracts at several major systems.

These agreements passed with strong member support:

• Mount Sinai Health System — ratified by 87% of voting nurses.
• Mount Sinai Morningside and West — ratified by 96%.
• Montefiore Health System — ratified by 86%.

Nurses at NewYork-Presbyterian also secured technology language after a historic 41-day strike, the longest at a major New York hospital in years.

These New York contracts carried the first-ever AI safeguards of their kind. The nurses won protections against systems where, according to our own State of Nursing survey, nurses were being asked to use AI tools they didn’t trust and were never trained on. The contracts are, in part, a direct response to that gap.

The second wave came from California, led by the California Nurses Association (CNA) and National Nurses United (NNU).

University of California nurses ratified a systemwide contract covering technology protections at UC campuses, including UC San Diego Health and UCSF Health. Because it’s one systemwide agreement, a single union announcement backs both medical centers. The contract empowered nurses to have a say in what AI technologies were chosen and how they would be deployed and later verified. 

Nurses at Sutter Davis Hospital also won AI language — and it was historic for a different reason. It was the hospital’s first-ever union contract, covering AI protections for roughly 250 registered nurses. (This AI language applies to Sutter Davis Hospital specifically, not to the wider Sutter Health system.)

The wave reached beyond California, too. Nurses with NNU and the National Nurses Organizing Committee bargained technology language into a multistate agreement with HCA Healthcare, covering 17 facilities across 6 states.

Strip away the legal wording and the technology protections in nursing AI contracts usually come down to three ideas:

1. AI can’t replace a nurse. A tool can suggest, flag, or draft, but it can’t stand in for a licensed nurse’s judgment or eliminate the role.
2. AI can’t be used to discipline nurses. An algorithm’s read on a nurse’s pace or charting can’t become the basis for a write-up or firing.
3. AI can’t drive staffing decisions. A model can’t quietly set how many nurses cover a unit, which is the lever that most directly affects patient safety.

Other language ensures that nurses get a say in what AI software is chosen to be used and deployed, and how it will be evaluated. Together, the language aims to give nurses a say in how these tools get rolled out, instead of finding out about a new system after it’s already live.

The short answer: trust is thin, and the stakes are clinical.

In a national survey from National Nurses United, 60% of nurses said they didn’t trust their employer to use AI in their best interest. Even more telling: 69% reported that AI-driven acuity tools didn’t match the actual care their patients needed.

When the tool measuring patient need gets it wrong, the nurse is the one left covering the gap. That’s the day-to-day reality pushing nurses to bargain — they want a brake they can actually pull.

The risk isn’t hypothetical. NYSNA says Montefiore recently shifted the duties of roughly a dozen utilization-review nurses to an AI system and eliminated their roles, and the union has filed a class-action grievance in response — this appears to be the first AI-attributed nurse layoff to prompt a formal union grievance. Montefiore has not publicly confirmed the union’s characterization, and Nurse.org is continuing to verify the details.

>>Listen to The Latest Nurse News Podcast

Here’s the set of systems we identified where nurses have ratified AI or technology language, by wave and state. As with the rest of Nursing AI Watch, these are the deployments we could confirm from the public record, not necessarily every contract that exists.

The early pattern suggests yes.

Two waves in two states, with strong ratification votes, give the next round of bargaining a template. When other unions sit down, they won’t be starting from scratch; they’ll be pointing at language that already passed.

There’s also a parallel track. While unions write private rules into contracts, lawmakers are writing public ones into statute. We track that side of the story in our coverage of AI legislation affecting nursing, and together, contracts and laws are starting to draw the same lines around how AI can be used in care.

Nurses in New York and California have done something rare: they’ve put hard limits on AI directly into their contracts. Across 7 systems, the message is consistent — AI can assist, but it can’t replace nurses, discipline them, or set staffing on its own. It’s an early pattern, but it’s a template the rest of the profession can build on.

Related Nursing AI Watch Analysis:

🤔 Does your contract say anything about AI? Share your experience in the comments below.

About the data: Nurse.org’s Nursing AI Watch is built from a structured dataset of AI deployments across 106 large health systems, refreshed each reporting cycle. The contract details here — the systems, unions, and ratification percentages — come from union press releases and public announcements, each cited inline. We count only what we can confirm from the public record, so the seven systems represent the contracts we could verify, not necessarily every nurse contract with AI language. When the evidence supports only a narrow statement, we scope the claim to match.

Nurse.org Analysis