Researchers in Hong Kong found most young screenwriters they surveyed had experimented with generative AI, mainly for ideas, but almost none found its dialogue or structure useful
Early Adopters
As Hollywood debates the role of artificial intelligence in writing, one of the few academic studies to examine how screenwriters actually use AI offers useful evidence. TheWrap’s Jackie Snow reported on the research in an article titled Chinese screenwriters are finding ways to use AI, is Hollywood ready to rewrite the script. The study, by researchers at the Hong Kong University of Science and Technology, was presented at the Conference on Human Factors in Computing Systems, a major academic gathering on human-computer interaction.
What The Researchers Found
The researchers surveyed 23 screenwriters aged 23 to 29 in mainland China. Eighteen of them had used generative AI in their writing process, mostly to spark ideas rather than to replace their own creative work. Lead author Yuying Tang told TheWrap that the writers were trying to understand the technology early.
The writers used tools such as ChatGPT for text and Midjourney for images. Uses included brainstorming, generating character backstories, sketching plot points, visualizing settings and getting unstuck when a story stalled. In one example, a writer created an image of a gorilla in a modern art museum, which inspired a story about animals in human society. The image did not write the story. It prompted the writer to imagine one.
Where AI Fell Short
The study’s findings on AI’s limitations are perhaps more telling than its findings on use. None of the six writers who tried generating dialogue found the output usable. Of the twelve who used AI for narrative structure, only two said they would do so again. Some outputs were described as stilted or formulaic.
That aligns with what many screenwriters report anecdotally. AI tools can generate plausible text quickly, but dialogue that reveals character, carries subtext and sounds like a specific person speaking in a specific moment is extraordinarily hard to produce. Structure, too, depends on intention: the writer’s sense of what a story means and how it should feel. Pattern-based systems tend to reproduce familiar shapes rather than surprising ones.
Expert Perspectives
Zorana Ivcevic Pringle of the Yale Center for Emotional Intelligence told TheWrap that AI is useful for brainstorming but weak at combining ideas into a coherent whole with nuanced emotion and social relationships. That distinction, between generating raw material and shaping it into meaning, is at the heart of what screenwriters do.
Van Robichaux, a television and animation writer who served on the Writers Guild of America’s AI working group before the 2023 strike, was more sceptical of the study itself, suggesting it overstated how widely generative tools are used. He noted that early-career writers are more likely to adopt AI and that AI’s pattern-based output tends toward averaging. He warned that the bigger risk is studios accepting mediocre, machine-assisted work.
The Hollywood Context
American studios are experimenting with AI in animation, visual effects and marketing, but the use of AI in screenwriting remains far more contentious. The 2023 WGA agreement established that AI is not a writer and that AI-generated material is not literary material under the contract. The 2026 agreement, according to reports on the tentative deal, added restrictions on using writers’ scripts to train AI. Those protections reflect writers’ concern that their work could be used to build systems that might later replace them.
A Tool, Not A Voice
The Hong Kong study suggests a nuanced picture. Young writers are curious about AI and willing to experiment. They find it useful in narrow ways, particularly for generating starting points. But they do not find it capable of doing the core work of screenwriting: creating characters who feel alive, dialogue that crackles and stories that mean something. Robichaux’s warning points to the real danger. The risk is not that AI will write great scripts, but that the industry might settle for average ones.
The study also raises questions about how AI might shape the next generation of writers. If early-career writers rely on AI for ideas, will they develop the same creative muscles as writers who generated every idea themselves? Some educators worry that heavy reliance on AI could weaken the struggle through which writers learn. Others argue that, used thoughtfully, AI could be one more tool, like a thesaurus or a writing prompt, that supports rather than replaces creative development. The answer will likely depend on how individual writers choose to use it.
Practical Guidance For Writers
For screenwriters considering AI tools, the research offers some practical lessons. If you use them at all, use them for what they appear to do best: prompting ideas, exploring possibilities and getting unstuck. Be cautious about relying on them for dialogue or structure, where the evidence suggests they add little. Understand the terms of any tool you use, including how your inputs may be stored or used. And remember that your value as a writer lies in your voice, judgment and lived experience, which no pattern-matching system can replicate.
The Takeaway
The study’s sample was small and focused on young writers in one country, so its findings should not be overgeneralized. But its central message resonates widely. AI may become a brainstorming partner for some writers. It is not, on current evidence, a substitute for the writer. The craft of turning ideas into stories that move audiences remains a deeply human one, and the screenwriters who master it will remain essential, whatever tools they choose to keep on their desks.