While the study’s conclusions are plausible, it’s crucial to note that individuals with an innate ability for invention may not require AI assistance to be innovative, according to Tuhin Chakrabarty, a computer science researcher at Columbia University specializing in AI and creativity, but not involved in this particular research.
While taking advantage of the mannequin’s support offers several benefits, there are also some potential downsides to consider. According to Chakrabarty, AI-generated narratives exhibit striking similarities in terms of semantics and content, with AI-written texts often betraying their artificial origins through characteristic features such as overly long, expositional sentences replete with stereotypical tropes.
“These types of anomalies probably also hinder overall creative potential,” he notes. Good writing isn’t just about conveying information, but rather skillfully showcasing ideas through subtle yet powerful storytelling. The AI is always speaking.
As AI-generated tales are limited to the data they were trained on, those created in the study lacked the distinctiveness of ideas conceived independently by human participants. If the publishing industry had chosen to adopt generative AI, the books we read might become increasingly uniform, since they would all be generated from models trained on the same dataset.
Because the implications of AI fashioning are far-reaching, it is essential to evaluate the capabilities and limitations of this technology as we navigate its impact on society and the economy, notes Professor Oliver Hauser of the University of Exeter Business School, a co-author of the study. “Just because knowledge has the potential to be transformative doesn’t mean it inevitably will be,” he notes.
