Rivalry and Collaboration Attitudes: NYU Study Finds Writers Need Both to Thrive in the Age of AI

When a screenwriter told New York University researchers last year that letting AI do her work would make her "miserable inside," she was onto something.

A follow-up study from NYU’s Tandon School of Engineering and Stern School of Business finds that the instinct to compete with generative AI, rather than simply embrace it, is associated with meaningful long-term benefits for writing professionals.

The catch: rivalry alone isn't enough either.

The 2026 study, led by Rama Adithya Varanasi, a postdoctoral researcher in Tandon's Technology, Management and Innovation Department, alongside Tandon Professor Oded Nov, and Batia Mishan Wiesenfeld, a professor of management at Stern, surveyed 403 professional writers across marketing, publishing, education, and the arts. Findings will be presented at the CHI Conference on Human Factors in Computing Systems this month.

The work extends a 2025 qualitative study by the same team, which interviewed 25 experienced writers and introduced the concept of "AI rivalry" — the idea that some writers proactively compete against AI rather than simply avoid it, targeting what they see as its weaknesses, such as its difficulty producing content rooted in specific communities or geographies.

The new research asked a larger question: what actually happens to writers' careers, skills, and satisfaction depending on how they orient themselves toward AI?

The study finds risks at both extremes. Writers who reported strong collaborative attitudes toward AI also reported higher short-term productivity and job satisfaction, but invested less in maintaining their own skills — the risk of over-reliance.

Writers who perceived AI as a rival reported stronger skill maintenance and greater investment in peer relationships, but that perception showed no significant association with productivity or satisfaction — the risk of under-reliance.

"The concern isn't that workers use AI," said Varanasi. "It's that they stop developing the capabilities that make humans irreplaceable. What this study tells managers is that they can't measure success purely by output. If the workflow removes the need for human judgment, the skill atrophies and that cost doesn't show up until it's too late."

Notably, rivalry attitudes didn't reflect a rejection of the technology. The data showed these writers reported more experience with generative AI than those who held neither orientation strongly. They studied the AI competition, rather than ignoring it.

The most striking result came from writers who scored high on both orientations simultaneously. This group showed the strongest associations with job crafting and skill maintenance across nearly every dimension measured, and posted productivity levels closer to the pure collaboration group — though satisfaction remained higher among pure collaborators — without sacrificing the long-term skill maintenance that pure collaborators showed less of.

"What surprised us is that rivalry and collaboration don't cancel each other out," said Wiesenfeld. "Writers who hold both orientations seem to use AI more deliberately. They get the productivity benefits without outsourcing the judgment."

The study is among the first to measure this tradeoff across a broad set of outcomes — relationships, tasks, cognition, skills, satisfaction, and productivity — drawing on expertise in both human-computer interaction and organizational behavior.

The implications for employers are direct. Organizations that push widespread AI adoption to boost efficiency may be optimizing for the wrong thing, particularly if those workflows come at the cost of workers practicing core human skills.

"Most organizations right now are still developing policies on how employees should relate to AI. " said Nov. "Our findings suggest that the relationship workers have with AI matters as much as whether they use it."

The researchers call for a new design structure that builds productive "friction" into AI tools, calibrating how much assistance is offered based on a user's reliance attitudes rather than defaulting to maximum engagement.

The team's next phase will test that concept directly. They are building prototypes of AI tools designed to promote appropriate reliance, and plan to expand the research beyond writing to other creative professions including game developers, graphic designers, and visual artists.

Funding for this research was provided by the National Science Foundation.