A hot topic during the recently-ended Hollywood writers’ strike was artificial intelligence. Apparently, we’ve reached the point at which the industry legitimately believes that artificial intelligence is capable of providing tangible benefit to screenwriting, and the industry is turning its head toward the possibilities of the future.
For instance, some have speculated that AI programs like ChatGPT could be trained to produce a first draft of a screenplay, which would then be scrutinized and redrafted by human writers. This, of course, poses a threat to the importance of those humans’ jobs, since, in such a scenario, they become relegated to proofreaders rather than the crucial originators of cinematic ideas. Their role in the creative process becomes significantly less valuable.
The question for viewers, though, is whether, if AI were involved, such a process would be truly “creative.” And the answer, to me, is fairly obvious: no.
ChatGPT and similar AI programs do not create, nor do they claim to. They function only as “learning” models that rely on feedback from humans to build proficiency in producing suitable language responses. In other words, they merely learn (with extreme efficiency) what their masters deem “good,” and they spit it out flawlessly.
Most viewers likely have a negative reaction to the notion that film dialogue and plot lines may be computer-generated in the near future. But I posit that large studios have already adopted an essentially AI-like approach to their work. Therefore, the dawning age of AI-written cinema is less a paradigm shift than a fulfillment of what mainstream audiences have already embraced.
Consider a notorious example of troubled blockbuster production: 2016’s Justice League. The Warner Bros. superhero film was originally directed by Zach Snyder, but in post-production, Snyder stepped down to grieve death of his daughter. Following his departure, Warner Bros. hired Joss Whedon, who had directed The Avengers (2012) and The Avengers: Age of Ultron (2015), films very different than Snyder’re prior releases. Whedon, at Warner Bros. behest, rewrote and reshot large portions of the film. His contributions had the predictable effect of fundamentally altering Snyder’s vision.
In all cases, Whedon’s modifications resulted in a more sanitized, safe product. We know this because in 2021, Snyder’s director’s cut was released, and it’s laughably different from the theatrical cut.
Of course, the reasoning behind Warner Bros. onboarding Whedon is easy to follow. Snyder’s previous DC-based films had met polarized responses. His characteristically dark and brooding style had turned off many viewers and critics, and his most recent output, Batman v Superman: Dawn of Justice (2016), was a full on flop. On the other hand, Whedon’s “Avengers” films had enjoyed enormous box office and critical success due in large part to his implementation of a light, comedic tone. Warner Bros. got cold feet after reviewing Snyder’s footage and brought in Whedon to save the film by Avenger-ifying it.
But I don’t care whether Warner Bros. makes logical corporate decisions. I’m interested in whether a movie has any creative value. If it doesn’t, why am I watching it?
Whatever you think of Snyder’s abilities—I certainly don’t claim to be a fan—the kind of studio meddling that affected Justice League is antithetical to anything resembling a “creative process.” That might seem obvious, but…did anyone care? Was anyone rankled that what they were watching was a blatant ripoff of the “Avengers” series? Were audiences offended that Whedon had near-exactly aped the plot of Avengers: Infinity War and had even digitally reimagined the villain, Steppenwolf, to visually resemble Thanos? Were they disturbed that Whedon had reworked the main expository sequence into a flagrant duplicate of the opening to Lord of the Rings?
Maybe some were. The reaction to Justice League was mixed, and the movie fell below box office expectations. Nevertheless, the film has an audience score of 67% on Rotten Tomatoes, and critics mainly called the film an improvement over most of its fellow DC-inspired movies—although still weaker than the Marvel movies that Whedon had tried so hard to copy. In summary: copy harder next time.
That brings us to artificial intelligence. AI, after all, is the king of copying. It can copy far better than Joss Whedon or any other human intelligence. Using its “learning” algorithms, it can do exactly what Warner Bros. hoped Whedon could: regurgitate components of previously successful content without indulging in risky stylistic ventures. It can process our tastes—or, more accurately, our purchasing patterns—and cater to them flawlessly.
Mainstream audiences have already proven willing to cede originality for familiarity, as long as it comes with the degree of quality control that Marvel, for example, has perfected. Will audiences, then, also be willing to sacrifice the literal possibility of originality, by seeking out movie and television content that has been, by definition, processed from the material they’ve already seen?
If they rebel, it will be only against the optics. The age of algorithm-generated screenwriting isn’t approaching: it’s already here. And it has been for quite some time.
–Jim Andersen
For more analysis, see my review of Oppenheimer.