What the WGA Actually Says About AI (And What It Means for You)
The flush of excitement that inhabits a writer getting their first story assignment can curdle quickly if they realize the brief they’ve been given includes AI-generated material. Does that disallow receiving writing credit? Is the project forever tainted? During the Writer’s strike in 2023, the WGA sought to place standards on AI use with the signatory companies they do business with.
The MBA that followed set up the WGA AI rules to govern these situations, but they only cover writers within the guild directly. This is Part 1 of a series on AI in screenwriting, and it’ll cover what’s in the Writers Guild of America’s 2023 MBA, and also what that means for writers who are not in the guild.
This article is general information about the industry’s relationship with AI, and should not be taken as legal advice—screenwriters with specific contract questions should consult a lawyer or the guild directly.
The headline rule
Under the 2023 Minimum Basic Agreement (MBA) between the Writers Guild of America (WGA) and the signatory companies of the Alliance of Motion Pictures and Television Producers (AMPTP), AI is not and cannot be credited as a “writer.” As such, AI-generated material is not considered “literary material” and can’t be counted as source material to reduce the writer’s credit or compensation. Literary material is any material (e.g. stories, treatments, plot outlines, screenplays, etc.) produced for use in production for television or film.
If AI-generated content could be counted as literary or source material, then it would be able to take the story credit from the actual writer. As per the WGA credit rules (which differs from copyright law that protects the work itself), the “written by” credit is a combination of the “story by” and the “screenplay by” credits. If AI-generated content could snatch the “story by” credit in lieu of an actual writer, then that would effectively nullify the “written by” credit as well. The 2023 MBA has prohibited this scenario altogether, rendering an AI screenwriting credit as nonexistent.
What this protects you from, specifically
The 2023 MBA agreement mainly deals with placing guardrails around generative AI (GAI). GAI is a subset of AI that uses generative models to create images, videos, and, most pertinent to the WGA, written material. Three core protections center the bulk of the 2023 MBA AI provision in regards to protecting writers. The first is as described in the example above–that your credit can’t be diminished by AI-generated pages a studio hands you. In addition to not affecting “story by” crediting, the AI-generated pages cannot affect the 33% and 50% threshold percentages that determine which writing credits a writer will or will not receive. Only writer’s contributions in relation to other writer’s contributions working on a project count toward those percentage thresholds that determine who gets which credit.
The next core tenet of these provisions is that writer’s cannot be required to use AI to perform their writing services. A writer can, if they wish, choose to do so with the consent of the company they’re working with as long as they notify the company and follow the company's AI policies while working on the project, but the agency lies with the writer. A studio cannot take punitive actions against a writer that refuses to implement AI tools or use generative AI material in rendering their services.
Furthermore, a company must disclose any AI-generated material or material that incorporates AI being given to a writer. So, if a writer decides to incorporate AI into their workflow and the company consents, the company still has to clearly label any AI material given to the writer as such.
The Writers Guild prioritized these immediate protections when negotiating the MBA, but they also acknowledged a future fight to be had that was effectively punted toward a future collective bargaining effort. “The WGA reserves the right to assert that exploitation of writers’ material to train AI is prohibited by MBA or other law.” This will be touched upon in the Unresolved section below.
The line most writers miss: when does this actually apply?
These protections are not the law of the land for all of screenwriting, however. Only the signatory companies in the AMPTP are a party to the agreement. Those companies include the Big Five (Disney, Sony, Universal, Paramount, and Warner Bros.), which will turn into the Big Four when Paramount and Warner Bros officially merge, the Big Tech streamers (Amazon, Apple, and Netflix), the big broadcast networks (ABC, CBS, Fox, and NBC), and some 300 other network and production companies. While 300 plus companies is quite the number, there are still many non-signatory companies writers can work with where the WGA’s 2023 MBA does not apply.
These protections activate at the point of employment or sale to a WGA signatory company. Anything work before the point of employment or sale to a signatory company does not fall under the purview of the MBA. Areas of the industry that aren’t covered include a spec script a writer is using AI to help draft, a competition entry, an indie feature that is financed outside of the studio system. As AI is a new and burgeoning technology, there are many gaps where its usage and restrictions are ill-defined. Said gaps should not be seen as a loophole to be championed and sought after, nor should writers fear working with non-signatory companies because these gaps still exist. Writers working in these areas of the industry need to plan with the understanding that there is no overseeing body governing the rules of non-signatory companies, and as such, writers should endeavor to understand the AI policies of any individual company they may engage in a working relationship with.
What's still unresolved
Circling back to the portion of the MBA stated above, “The WGA reserves the right to assert that exploitation of writers’ material to train AI is prohibited by MBA or other law,” is a concession at the same time as it is a promise (whether it turns out to be an empty promise remains to be seen). The concession is that the guild could not come to specific and concrete terms with AMPTP on the issue of writers’ materials being used to train AI. The promise is this vague statement that they reserve the right to assert so in the future. Many lawsuits are being litigated on the issue of training data, which will help clarify how future bargaining agreements can broach the subject.
In 2025, the WGA West (WGAW) and WGA East (WGAE) joined in filing an amicus brief in Reuters v. Ross Intelligence, in support of the plaintiff’s argument that AI training that uses copyrighted material without consent does not fall under Fair Use doctrine. The WGAW also endorsed the Protecting Consumer From Deceptive AI Act which would require generative AI content that creates or modifies audio or visual content to be disclosed as AI. The legislative battle regarding the training and use of AI will be ongoing for the foreseeable future. You can expect to see this battle play out on screen as well as in court, with the ethics of AI becoming plot points or allegories for filmmakers to explore.
Another point of ambiguity in the 2023 MBA is the acknowledgement that writers can be required to use AI is practically very difficult to enforce when AI tools are ubiquitous in much of the software writers now use. The flip side also proves true for the signatory companies–how could they know if a writer used AI without their knowledge if it is not directly identifiable in the actual text they receive?
AI is still the wild west in terms of litigation. The technology’s growth outpaces the judicial system's ability to adjudicate the issues inherent to AI promptly. It will take some time for the dust to settle.
What this means for you right now
It’s important for any non-signatory writer to understand the rules now so they’re not caught flat-footed later. Everyone in the industry will have a different personal standard. Some will openly embrace a new technology, and champion anything to help catalyze their creative process. Others will recoil from something they consider inhuman and not creative in the slightest. Consider setting your own personal standard for AI use and disclosure even when there’s no contract requiring it. Reputational standards in the industry often precede anything else. In addition to understanding the rules, it could be illuminating to also look up the views of your favorite filmmakers and try to understand their perspective. If you look up ten filmmakers' views, are they mostly aligned? Do they differ quite substantially? This is a time of flux in the industry. And times of flux bring ambiguity.
A decade before AI became a live labor issue within the industry, Spike Jonze wrote a film about human-AI relationships in Her. Do you have any story ideas about where our relationship with AI will be in the 2030s?
