Writers’ Strike Brings Generative AI to Hollywood

Hollywood’s writers’ strike, which began on May 2, has turned generative artificial intelligence into a labor bargaining issue. The Writers Guild of America (WGA) wants to set limits before tools such as ChatGPT become fully integrated into the production pipeline for television series and films.

It is not the strike’s main cause: streaming compensation and the size of writers’ rooms account for much of the dispute. But AI has entered the negotiations because it could change who does the writing, how a work is credited, and how much value human labor retains in a copyright-based industry.

The proposal: AI should not write or rewrite scripts

The Writers Guild of America (WGA), which represents about 11,500 film and television writers, has proposed banning the use of AI to write or rewrite literary material covered by its collective bargaining agreement. It has also called for machine-generated text not to be considered source material, a distinction that matters for credits, payments, and subsequent rights.

In practice, the union wants to prevent a situation in which a studio gives a writer an AI-generated draft and asks them to revise it for less pay than an original script would command. The debate is not just about whether a machine can produce dialogue or an outline. It is about whether that text becomes the contractual starting point for a work and weakens the professional standing of the person who turns it into a usable script.

The WGA has also proposed barring studios from using scripts covered by the agreement to train AI models. These systems learn patterns by processing vast collections of text; for writers, allowing that use without rules would turn their accumulated work into raw material for a tool that could compete with them.

Studios agree to talk, but not to add limits to the agreement

The Alliance of Motion Picture and Television Producers (AMPTP), which negotiates on behalf of major studios and platforms, rejected the WGA’s specific AI proposals. Instead, it offered annual meetings to discuss technological advances.

The gap between the two positions is significant. Regular meetings would allow the parties to track the technology’s evolution, but they would not create a contractual obligation governing what companies can do with it. The union is seeking the opposite: rules that apply now, before decisions about automation become normalized in writers’ rooms.

The disagreement comes just months after the mass adoption of assistants capable of generating text from prompts. ChatGPT, released by OpenAI in late 2022, and subsequent models have made a capability visible that until recently seemed experimental: producing synopses, scenes, dialogue, and new versions of a text within seconds.

That does not mean these systems can create a television series or film on their own. A script requires narrative decisions, an understanding of the assignment, revisions, and creative accountability. Models can also make up facts, repeat formulas, and produce inconsistent text. But they can automate parts of the process, and that possibility alone is enough to shift the balance of power among studios, producers, and writers.

A problem of credits, pay, and intellectual property

The AI debate affects several mechanisms that underpin the profession. Writing credits determine professional standing and pay; treating a work as original material affects contracts and rights; and residuals—later payments tied to the exploitation of films and series—make up a central part of many writers’ income.

If a company presents automatically generated text as the starting point, it will have to decide who is responsible for it, how human contributions are recognized, and what happens when the system has been trained on protected works. These are still open legal questions in the United States, but the labor dispute offers another way to address them: rather than waiting for the courts to define every limit, negotiate them in the collective agreement.

Hollywood has already gone through technological transformations that changed its labor rules, from the arrival of sound to cable television and streaming. The difference is that AI does not merely change the distribution channel or editing tools: it intervenes in the text itself, which is at the core of screenwriters’ work.

The 2023 negotiations will be an early test of how far collective bargaining agreements can impose conditions on the use of generative models. Whatever is agreed—or left out of the deal—will be watched by other creative sectors facing the same dilemma: adopting a new tool without turning their professionals’ work into an uncontrolled, uncompensated resource.