As artificial intelligence (AI) and its capabilities continue to progress, writers and others have expressed concerns over its effects on the media and publishing industries. According to the Harvard Business Review, the ‘creative content’ industry, valued at a staggering $14 billion per year, is at risk of being fundamentally reshaped by AI.
News agencies and magazines have already laid off hundreds of writers, turning to AI to create articles and catchy titles and attract more paying readers. This has created significant issues for anyone in this industry as they work to create unique and engaging stories that stand out compared to AI-created content.
While generative AI has been seen with some trepidation by writers over the last couple of years, by some measures, it may also be more helpful than expected. A team of researchers from the University of Exeter Business School, the Institute for Data Science and Artificial Intelligence, and the UCL School of Management found that generative AI boosted an author’s creative work and suggested changes that would make the story more engaging to the target audience. The findings were published in Science Advances.
“Our results provide insight into how generative AI can enhance creativity, and removes any disadvantage or advantage based on the writers’ inherent creativity,” explained Oliver Hauser, Professor of Economics at the University of Exeter Business School and Deputy Director of the Institute for Data Science and Artificial Intelligence in a recent press release.
Short Story Enhancements
To understand how AI affects the writing and creativity processes, the researchers asked 300 participants to write a micro-story of only eight sentences for a target of young adults with varying access to AI. The writers were split into three groups: one without access to AI, one using ChatGPT to give a three-sentence starting idea, and a third group allowed to use up to five AI-generated sentences as inspiration for their story.
Then, the scientists asked a team of 600 people to judge the micro-stories, assessing their uniqueness or novelty, whether they fit the required target audience, and whether they were engaging enough to be further developed and later published.
The researchers found that those judged to produce “less creative” stories received significant boosts from AI assistance, including a 26.6% increase in better writing and a 15.2% increase in being more engaging.
The researchers found similar trends across all three groups, with those using AI the most and those using second most scoring 8.1% and 9% higher for uniqueness, respectively. The group that used AI the most had micro-stories that were the funniest, most enjoyable, and least boring.
Other AI Creativity Evaluations
Beyond examining how a general audience judged AI’s effectiveness, the researchers also used a separate metric to study the writers’ creativity with and without AI. The researchers implemented a Divergent Association Task (DAT), a psychological test that measures a person’s creativity. Using this metric, the researchers found that those who scored the highest on a DAT benefitted the least from using AI-generated ideas. Those with lower DAT scores benefitted more significantly, putting them on a similar level to writers with higher DAT scores, which the researchers found equalized creativity between more and less creative writers.
Similarly, the researchers analyzed the similarities between AI-generated stories to see their uniqueness in using OpenAI’s embedding application programming interface (API). The team found a 10.7% increase in similarity between stories whose writers used AI-generated ideas compared to those who didn’t.
“While these results point to an increase in individual creativity, there is risk of losing collective novelty,” added Anil Doshi, Assistant Professor at the UCL School of Management in the press release. “If the publishing industry were to embrace more generative AI-inspired stories, our findings suggest that the stories would become less unique in aggregate and more similar to each other.”
A Word of Warning
While AI can boost creativity for writers, and effectively help with issues such as writer’s block, the researchers of this study offer words of warning in overusing generative AI for creating content.
“In short, our results suggest that despite the enhancement effect that generative AI had on individual creativity, there may be a cautionary note if generative AI were adopted more widely for creative tasks,” Hauser added.
Kenna Hughes-Castleberry is the Science Communicator at JILA (a world-leading physics research institute) and a science writer at The Debrief. Follow and connect with her on X or contact her via email at kenna@thedebrief.org