'Gladiator II' Writer David Scarpa: Why AI Won't Kill the Hollywood Screenwriter

David Scarpa, who wrote "Gladiator II" and "Napoleon" for director Ridley Scott, published a direct counterargument to AI doom narratives in The Hollywood Reporter on May 1, 2026. His core claim: audiences care about what human beings achieve, not what machines can produce, and that distinction will keep professional screenwriters employed.

The Moment That Started the Debate

On February 11, 2026, "Deadpool and Wolverine" writer Rhett Reese watched a viral AI video of Tom Cruise fighting Brad Pitt, created from a two line prompt in ByteDance's Seedance 2.0 model. Reese posted his reaction publicly: "It's likely over for us". His full statement: "In next to no time, one person is going to be able to sit at a computer and create a movie indistinguishable from what Hollywood now releases".

Scarpa read it differently. The capability to generate a film and the demand to watch one are not the same thing, and the history of technology offers a precedent.

The Chess Precedent

IBM's Deep Blue defeated world chess champion Garry Kasparov in 1997. Computers now surpass human chess ability by a margin that makes direct comparison meaningless. Human competitive chess has thrived globally in the decades since. AI engine tournaments, which computers dominate completely, remain a technical curiosity. "Nobody cares about them", Scarpa wrote.

The analogy is precise. The question for screenwriters is not whether AI can generate a feature film script. It is whether audiences will seek out a film made without human authorship. Scarpa's answer: they will not. "Human beings are social animals. We care about what other human beings do". On AI generated movies specifically: "The question is whether anyone will watch it".

Three Writers, Three Predictions

Scarpa's position is the least covered of the three dominant stances among major Hollywood writers. Reese's pessimism ("it's likely over for us") and Paul Schrader's accommodation (the "Taxi Driver" writer has said he is ready to make the "first AI movie" if the right script arrived) both generated headlines. David Simon, creator of "The Wire", has stated publicly that AI represents an existential threat to writers' craft. Scarpa is the one arguing that audience behavior, not contract language, is the structural protection.

The contrast matters because all four writers are operating at the highest levels of the industry. They are not reacting to a hypothetical. They are describing what they expect from tools they have already encountered in their working environments.

What the Industry Has Done in Parallel

The WGA four year deal signed in April 2026 secured AI licensing language: studios cannot use writers' scripts to train AI models without compensation. The deal did not fully resolve the replacement question, since studios face no payment obligation if AI is trained on generically similar material rather than specific scripts. The labor contract and Scarpa's cultural argument are two separate, complementary responses to the same threat.

The Academy's 99th Oscar rules also address AI authorship: scripts generated primarily by AI are ineligible for screenplay nominations. The institution established a floor for what counts as human authorship at the same time Scarpa was making his cultural argument in the press. Neither the contract nor the rule changes address Scarpa's underlying point. His argument is not about protection. It is about preference: that audiences will choose human stories over AI generated ones the same way chess fans still tune in for human tournaments after Deep Blue.

Filmmakers who want to examine the current state of AI video generation, the capability side of the debate Scarpa is describing, can work with the tools directly in AI FILMS Studio's video workspace.

What Reese's Viral Video Actually Showed

Rhett Reese's February 11, 2026 response to the ByteDance Seedance 2.0 clip was specific about what alarmed him. He was not objecting to AI tools in general. He was reacting to a demonstration that a single person could generate a Tom Cruise and Brad Pitt action sequence from a two line prompt, indistinguishable in visual quality from a studio production.

His "it's likely over for us" statement described a specific capability threshold he believed he had just watched crossed. The question Scarpa is responding to is whether the existence of that capability changes the audience demand for human written films. Reese said yes. Scarpa said no.

Both writers are reacting to the same video. The disagreement is about what it means, not about what it showed.

Deep Blue, Kasparov, and the Chess Analogy in Detail

IBM's Deep Blue defeated Garry Kasparov in 1997. The defeat was not the end of human chess. It was the beginning of a period in which human competitive chess grew globally, with grandmaster counts rising, prize pools increasing, and streaming audiences for elite tournaments reaching new highs.

AI chess engines now calculate billions of positions per second and operate at Elo ratings no human can approach. Human competitive chess and AI engine tournaments are treated as separate categories. The human tournaments draw audiences. The engine tournaments do not.

Scarpa's analogy identifies the variable that determines whether audience preference survives capability. Chess fans attend human tournaments because they are watching what a specific human mind can do under pressure, not because they cannot access a stronger player. The same logic applies to film audiences, in his argument.

David Simon and the Opposite Conclusion

David Simon, creator of "The Wire," has stated publicly that AI represents an existential threat to writers' craft. His argument is not about audience preference. It is about the economic infrastructure that supports professional writers.

If studios can generate scripts that are "good enough" for certain production categories at lower cost, the number of professional writing jobs available decreases regardless of whether audiences prefer human written films. Simon is describing a labor market argument, not a cultural one.

Scarpa's cultural argument and Simon's labor argument are not directly opposed. Both could be true simultaneously: audiences prefer human written films, and the writing profession still shrinks because studios can capture a share of the market with lower quality AI generated content and a reduced professional writing budget.

Scarpa's Work as Evidence for His Own Argument

"Gladiator II" grossed over $460 million worldwide. "Napoleon" grossed over $220 million. Both films required screenwriters who could construct historical epic narratives with the cognitive complexity that makes them work dramatically and commercially.

Scarpa's argument that audiences care about human achievement is grounded in productions that tested that claim at commercial scale. His films performed because they contained the kind of narrative construction he is arguing AI cannot reliably replicate.

That does not prove his argument for all genres and all audiences. It proves it for the specific kind of large scale, historically grounded drama he writes. His confidence in the argument reflects direct experience of audiences choosing his work at theatrical scale.

The Audience Preference Gap

A March 2026 study by Alvarez and Marsal of nearly 2,000 US consumers found that only 33 percent believe AI effectively crafts emotionally resonant content. Among industry creators and marketers, 77 to 78 percent believe it does. That 40-plus point gap is the empirical record behind the argument Scarpa is making in cultural terms.

The industry believes AI can do the work. The audience is not convinced. Scarpa's argument is that the audience's conviction is more durable than the industry's optimism, and that the cultural reasons for audience preference will not dissolve as the technology improves.

His essay does not cite that data. He does not need to. His argument is constructed from first principles about what audiences are watching for when they watch a film.

What the Essay's Publication Timing Signals

Scarpa published the essay in The Hollywood Reporter on May 1, 2026. The timing placed it within weeks of the WGA's four year AI deal, one month before the Academy announced its 99th Oscar authorship rules, and two months before the first fully AI generated feature premiered at Tribeca.

The essay arrived at the moment when every institutional actor in Hollywood was making a specific bet about what AI in filmmaking means. Studios were integrating AI into production workflows. Guilds were negotiating contract protections. The Academy was writing authorship rules. Scarpa published a cultural argument that bypassed all of those institutional responses and addressed the question he considered primary: whether audiences will want the output.

What "Gladiator II" and "Napoleon" Required of Their Writer

"Gladiator II" and "Napoleon" are both historical spectacles that require a screenwriter to construct a coherent dramatic narrative from documented historical events involving real political figures. That task requires not only narrative craft but research, judgment about which documented facts support dramatic compression, and decisions about where historical record ends and invention begins.

Scarpa's argument that AI cannot replace the professional screenwriter is grounded in his experience of what that specific kind of writing requires. Writing a historical epic for Ridley Scott, a director with specific and exacting demands about historical authenticity and dramatic structure, is not the same as generating a script from a prompt.

Whether AI can produce a first draft that a skilled screenwriter could then develop is a different question from whether AI can produce the final script independently. Scarpa's essay addresses the second question. The first is a separate proposition that neither confirms nor refutes his core argument.

The Dreams of Violets Test Case

The first fully AI generated feature, Dreams of Violets, premiered at Tribeca's official selection in June 2026, approximately six weeks after Scarpa published his essay. The film answered the question of whether an AI generated feature could reach a major festival. It did not answer the question Scarpa is asking.

Tribeca's official selection is an artistic evaluation by a programmer making curatorial choices. The commercial box office test Scarpa identifies as the real threshold is a different evaluation by a general audience making a consumption choice.

Dreams of Violets at Tribeca is evidence for the AI side of the capability debate. Whether it generates the commercial result that changes studio investment decisions is the question Scarpa said matters. That result will not come from Tribeca.

Scarpa's chess analogy established the benchmark clearly enough that when the commercial result does arrive, the industry will recognize it against the threshold he named. That is his essay's lasting contribution to the debate, regardless of which side is eventually proven right.

---

Sources

The Hollywood Reporter | Variety