While studios and unions debate long -term implications of artificial intelligence, the writers at the start of their career in China find ways to incorporate tools into their creative process, according to a recent study by Hong Kong. The study of the University of Sciences and Technology revealed that most participants use tools like Chatgpt and Midjourney to think about ideas, build characters and take off in a creative way.
Presented this spring at the 2025 conference on human factors in computer systems, the study studied 23 screenwriters, aged 23 to 29, in mainland China. Eighteen said they had used a generative AI as part of their writing process – not to replace creativity, but to restart it.
The study seems to be the only university research to date that closely examines how the writers use AI. The fact that he comes from researchers in China – not in the United States, where AI is experienced in fields such as animation, VFX and marketing – can reflect Hollywood’s particular discomfort to apply technology to the scenario. Some fearing the AI generator applied to the narration will only do only mediocrity – and will cut human creators – in Hollywood.
In the Chinese study, writers described the use of AI tools to generate characters’ stories, draw points of the plot or view the parameters. A participant said they used Midjourney to generate an image of a terrified gorilla standing in a museum of modern art. This unexpected combination has inspired a speculative history on animals sailing on human society. Others have described moments of similar surprise – when AI has surfaced a detail or a juxtaposition in which they would not have arrived alone.
“Most of them are positive, even if some will have concerns about replacing their work,” said Yuying Tang, the main author of the study looking for how AI systems intersect with creative fields such as cinema, design and fine arts. “They also recognize that AI will probably improve over time.”
Tang added that participants considered their experimentation with AI as a means of preparing future changes in the creative landscape.
“They try to understand it early,” she said. “Sometimes they do not directly use the release, but that gives them ideas.”
The usefulness of AI at this stage can reside in the way it imitates the functioning of human inspiration – causing loose associations that lead to new ideas, according to Zorana Ivcevic Pringle, a main researcher at the Yale Center for Emotional Intelligence. This is why generative tools can help think about a brainstorming, but to fail regarding more advanced tasks.
“Where AI is not useful, I break different ideas and create a set coherent with a nuanced understanding of emotions and social relationships,” she said. “We start with a coherent project and iters until he comes to life.”
The study supports this division. While the participants turned to AI to break the creative blocks or generate raw materials, few found the tools useful for dialogue or structure. None of the six writers who tried to generate a dialogue said that the results were usable. Of the 12 who used AI to shape the narrative structure, only two said they would do it again. In some cases, AI’s responses have been described as stuck or formulated – the type of outing that could pass for the content, but not for crafts.
Tang said that if the study focused on Chinese writers, the underlying creative challenges are similar elsewhere. “I don’t think there is a huge difference between different places,” she said, “the needs are similar.”
Van Robichaux, a television and animation writer who sat on the working group on the AI of the Writers of America leading to the 2023 strike, told Thewrap that the study had done a solid job to capture the structure of the scenario, but overexpressed how much the generative tools are really used. He felt that the authors were based too much on academic quotes and indicated how the study framed the WGA strike as an example. “The concern was more where things were directed than where they were,” he said. “It was not a reaction to generalized use.”
However, he said that the conclusions could be relevant to Hollywood. The fundamental principles of scripting – from development to revision – are largely consistent through borders. Rather than the location of the participants, Robichaux said that their career status had probably shaped their opening to AI.
“Young people would be more willing to implement these things in their workflow,” he said. “But you quickly learn about things that work as they work for a reason.”
He added that many ways in which participants used AI, as for the generation of characters or the guests of history, do not reflect how more experienced writers generally work. Having worked on a project in China, he said that he would expect the distinction to also.
While Robichaux said that he was not opposed to AI, he stresses that the operation of these tools – from models in existing data – is fundamentally in contradiction with what writers often try to do: break the formulas, not reproduce them.
“One of the biggest concerns of the content generated by AI is this average,” he said. “He strikes this” fairly good “bar but is quite flat. We have already evolved towards this by ourselves without increase. ”
He sees the greatest risk not in the AI to replace writers, but in the studios becoming comfortable with mediocre and machine -assisted work. “Hollywood is already struggling with a lot of medium narration,” he said. “We don’t need the pushing ia even further in this area.”
