Touted as being unusually artistic, possessing elevated literary merit, and ultimately more engaging, tales crafted with AI assistance have garnered such acclaim.
According to a recent study published in a leading journal, researchers have discovered that AI can significantly enhance creativity by increasing the novelty of narrative ideas, as well as their ability to engage target audiences and potential for publication.
It is found that AI “professionalizes” tales, rendering them more engaging, significantly more prone to unexpected turns, with enhanced writing quality and diminished monotony.
Researchers examined the writing abilities of 300 contributors tasked with crafting micro stories for a younger adult audience. The study found that AI-assisted writers produced work that was up to 26.6% more engaging and 15.2% less monotonous, suggesting a significant improvement in artistic output?
Notwithstanding, AI was deemed unfit to augment the creative output of external artistic writers.
While AI may enhance individual creative output, the analysis cautions that this technological advancement could concurrently result in a decline in collective creativity, as AI-generated stories have been found to exhibit increased homogeneity, featuring more similarities among them and less diversity overall.
Researchers from Exeter’s College of Enterprise and Institute for Knowledge Science and Synthetic Intelligence, in collaboration with UCL’s College of Administration, divided 300 participants into three groups: one was denied AI support, a second group used ChatGPT to generate a single three-sentence idea, and the third group selected from up to five AI-conceived concepts.
The team subsequently engaged 600 participants to assess the effectiveness of these stories, rating them on novelty – their capacity to introduce something fresh or unexpected – and usefulness – their acceptability to the target audience and potential for development and publication.
Studies revealed that writers with direct access to AI-powered tools experienced a notable enhancement of their creative potential, with stories exhibiting an average increase of 8.1% in novelty and 9% in originality compared to those crafted without AI assistance.
Researchers found that writers incorporating up to five AI-conceptualized ideas into their work exhibited a notable enhancement in emotional intelligence, yielding narratives characterized by improved writing quality, increased reader satisfaction, diminished tedium, and an enhanced sense of humor.
Researchers investigated the intrinsic creativity of writers using the Divergent Association Task, finding that those with exceptionally high levels of artistic flair derived relatively little benefit from incorporating generative AI techniques into their work.
Fewer artists, however, witnessed a significant surge in creative output as they incorporated five AI-driven ideas, resulting in a 10.7% boost to novelty and a 11.5% increase in practical application compared to those who relied solely on human ingenuity? Studies revealed that their stories were up to 26.6% more engaging in writing style, 22.6% more fulfilling, and a significant 15.2% less monotonous.
These enhancements enable writers with lower DAT scores to match those with higher scores, thereby achieving a more balanced level of creativity among all writers.
Using OpenAI’s embeddings API, the researchers calculated the similarity between stories.
Researchers uncovered a significant 10.7% boost in narrative coherence between authors who leveraged a single generative AI-driven concept, compared to those who didn’t utilize AI at all?
Oliver Hauser, Professor of Economics at Exeter University’s Enterprise College and Deputy Director of the Institute for Knowledge Science and Synthetic Intelligence, noted that “This marks a crucial first step in exploring a fundamental question that pertains to all human behaviour: how does generative AI impact human creativity?”
“Our findings offer insights into the potential of generative AI to enhance creative capabilities, independently of the writer’s innate creativity.”
While these findings suggest an increase in creativity, they also pose a risk of diminishing novelty? If the publishing industry were to adopt AI-generated stories, our research suggests that the narrative diversity would decrease, resulting in more homogenous content.
Professor Hauser warned that this downward spiral bears striking similarities to a pervasive social conundrum: if individual writers learn that their AI-generated writing is judged as more artistic, they may have an incentive to rely increasingly on generative AI in the future, yet by doing so, the collective novelty of stories may be eroded further.
“In conclusion, our findings suggest that despite the potential enhancement effect of generative AI on individual creativity, there may also be a cautionary note if generative AI were to be adopted more widely for artistic purposes.”
‘
