Study reveals how writers compete with AI

Writers are not passive victims of AI disruption but active crafters of their professional futures, according to new research from New York University presented at the ACM CHI conference on Human Factors in Computing Systems in Yokohama, Japan.

Generative AI technologies are transforming the writing profession, eliciting opposing reactions. “If I let AI do my work, it would make me miserable inside,” one screenwriter told researchers, while a paralegal countered, “Individuals who are not using it are at a serious disadvantage.”

The study was overseen by NYU Tandon’s Oded Nov (Morton L. Topfer Professor of Technology Management) with lead author Rama Adithya Varanasi (Postdoctoral Researcher at NYU Tandon) and Batia Mishan Wiesenfeld (Professor of Management at NYU Stern). The team interviewed 25 professional writers with an average of 17 years of experience, all with at least 12 months of exposure to generative AI.

Using “job crafting” theory, which examines how workers redesign their roles to derive greater meaning from their work, the researchers identified four distinct strategies that writing professionals employ in response to generative AI—two focused on resistance and two on adoption.

1. Human-driven expansion (AI Resisters): Writers strengthen their identity by making their human labor visible, building personal brands and forming collaborations.
2. Human-driven localization (AI Resisters): Writers create niche identities, appeal to selective audiences that value human work, and sometimes reduce quality to compete with AI’s speed.
3. Generative AI-driven expansion (AI Adopters): Writers use AI to enhance creative workflows, generate alternatives, overcome blocks, and assist with challenging communications.
4. Generative AI-driven delegation (AI Adopters): Writers offload tedious tasks to AI, reduce emotional labor, and minimize dependencies on colleagues.

The research uncovered significant evidence of “AI rivalry” among some resistors—professionals actively competing against AI rather than simply avoiding it. Writers employing resistance strategies deliberately target what they perceive as AI’s weaknesses, such as its limited ability to generate content specific to a geographic area, community or context.

“I’m taking steps to be more independent,” one SEO writer shared. “I have started 11 websites. The baking website is my best contender.” Many writers are creating revenue streams where they own the content, anticipating industry disruption.

Meanwhile, adopters engage in significant “AI managerial labor”—the invisible work of designing prompts, cleaning outputs, and verifying results. This requires substantial workflow changes, as one paralegal noted: “With ChatGPT, I need to block two to three hours to complete the prompting… If I leave it midway… it is extremely difficult to follow the reasoning.”

Economic implications loom large, as some senior professionals eliminated dependencies on junior writers. “The difference now is that I’m not dealing with a lot of writers; I’m not giving them therapy sessions,” said one publishing house owner who replaced staff with ChatGPT.

A key distinction emerged: Resisters shape both identities and practices, while adopters focus primarily on practices without significant identity work. This stems from AI’s anthropomorphizing features, which some writers perceive as competing with their creative identities.

Varanasi emphasized: “Resistors engage holistically in their human potential to shape both identity and practices… We introduced the notion of AI rivalry to show resistors engaged in constructive strategies while viewing generative AI as a rival similar to how they would treat a human rival.”

The researchers recommend creating communities where resisters and embracers can share insights, as both groups would benefit from understanding each other’s approaches to navigating this technological transformation.