The First AI Labor War Has Been Won (By Workers)
The AI labor war did not begin with a government regulation or a corporate policy announcement. It began on a picket line in Hollywood in May 2023, when 11,500 screenwriters walked off their jobs and made AI protections specifically, restrictions on studios using AI to replace writers or use AI-generated scripts with human writers 'punching up' the content a central bargaining demand. One hundred and forty-eight days later, they ratified a contract that contained what many consider the first major union-management agreement regulating the role of AI across an entire industry. The strike lasted four months and ended with workers winning on the specific AI protections they demanded. The frame that 'workers have no power against AI adoption' was the first thing that fell. Here is what the first AI labor war achieved and what it means for every professional trying to understand where the balance of power sits in 2026.
In 2023, Hollywood writers and actors went on strike to win AI protections. They won. In 2025, nurses won contract language restricting AI-driven clinical decisions. In 2026, the ProPublica newsroom staged a 24-hour strike over AI job protections. Workers are not waiting for policy. They are negotiating directly and winning.
What Hollywood Workers Won
The Writers Guild of America contract, ratified in October 2023 after 148 days on strike, established three protections that set the template for AI labor negotiations across industries. First, AI cannot be considered a 'writer' its output cannot constitute 'literary material' and cannot be used to determine credit, compensation, or the terms of a writing engagement. Second, a studio can provide AI-generated material to a writer as a reference, but cannot require the writer to use it, and this use cannot affect the writer's compensation. Third, the studio is required to disclose if any material provided to a writer was produced by AI, giving writers the information they need to protect their credit and compensation rights.The SAG-AFTRA contract, ratified after 118 days on strike, contained more extensive protections reflecting the more immediate threat AI posed to performers. Studios must obtain performers' explicit consent before creating a digital replica of their face, voice, or likeness. When a digital replica is used, the performer must be compensated at a rate comparable to what they would have received if they had been on set. These protections address the specific threat that the technology was already capable of delivering in 2023 not a theoretical future concern but an active industry practice that had begun without consent or compensation.
The Wins That Followed in Other Industries
Healthcare: New York City nurses, 2025
15,000 New York City nurses went on strike and won contract language restricting specific AI misuse in clinical settings including restrictions on AI tools that override nursing clinical judgment, that influence patient prioritisation in ways prioritising cost over care, or that generate automatic performance assessments and discipline without human review. The National Nurses Union, representing 100,000 members, made AI governance a primary bargaining demand across 2025 contract cycles. Their position that AI in clinical settings without nurse input is a patient safety issue, not just a labour issue proved compelling to the public and to hospital management.
Logistics: UPS Teamsters
The Teamsters negotiated contract language that became a model for technology governance in labour contracts. Article 6, Section 4 of the UPS agreement established a joint committee to review any 'meaningful change in equipment or materials' defined broadly enough to include AI systems affecting job scope, workload, wages, or headcount. UPS must notify the union in advance of any qualifying change. Critically, if AI creates new work that replaces or modifies existing bargaining unit work, bargaining unit employees perform that new work. The principle established: AI-driven efficiency gains do not automatically translate to contractor replacement or headcount reduction without negotiation.
Casino workers: Culinary Workers Union, 2023
Las Vegas casino workers negotiated what became one of the most quoted AI labour contract provisions: a severance package of $2,000 for each year of service if an employee's role is eliminated due to 'technology or AI.' This established a financial accountability principle that compensation for AI displacement should be proportional to worker tenure not a flat exit package that ignores contribution history.
Journalism: ProPublica and the AP, 2026
In April 2026, ProPublica's union staged a 24-hour strike over AI job protections and compensation structures as AI increasingly influenced editorial workflows. The same week, union leaders at The New York Times sent leadership a letter calling the company's AI policies 'woefully inadequate' after an AI-generated book review was found to contain language similar to a Guardian review. The Associated Press union alleged management had refused to bargain over AI deployment. Across the journalism industry, the fight is specifically about two things: whether AI-generated content can substitute for union members' work, and whether the data journalists have produced which is being used to train AI models entitles them to compensation for that use.
What Workers Are Negotiating For
- Consent before deployment: no AI system that affects job scope, monitoring, performance evaluation, or compensation can be introduced without advance notice and a negotiation period
- Credit protection: AI cannot replace human authorship credit for creative work writing, design, journalism, or performance and AI-generated outputs cannot be attributed to human workers without their knowledge
- Severance for AI displacement proportional to tenure, not a flat exit package
- Human review requirements for algorithmic discipline, performance scoring, and termination decisions automated systems cannot be the sole basis for employment consequences
- Likeness and data rights: workers' faces, voices, behavioral data, and creative outputs cannot be used to train AI models without explicit consent and compensation
- Transparency: workers must be informed when AI systems are being used in decisions that affect their employment, compensation, or working conditions
What This Means for the Broader Balance of Power
The AFL-CIO's Workers First AI Summit in late 2025 produced a coordinated legislative and bargaining strategy that reflects how much the landscape has changed from 2023. In 2023, individual unions were negotiating AI protections as novel contract provisions. By 2026, the AFL-CIO has a formal AI agenda, is coordinating across industries, and is pushing for legislative protections in state legislatures where federal action is unavailable under the current regulatory environment.The frame that workers are passive recipients of AI adoption that corporate decisions about AI deployment are natural forces rather than negotiable choices has been substantially damaged by three years of successful collective action. The Hollywood wins proved that AI protections could be won in contract negotiations. The healthcare wins proved the principle extended beyond creative industries. The casino severance provision proved that financial accountability for AI displacement could be written into contracts. The 2026 journalism strikes are testing whether these principles extend to workplaces where AI is more directly substituting for the specific work that union members do. The first AI labor war is not over. But the workers who fought its first battles won more than most observers predicted and the victories they achieved are shaping what is negotiable in every contract cycle that follows.