ChatGPT's Short Stories Are Outshining Human Writing

AI may not be ready to pen the next Odyssey, but readers already seem to like what it's writing. A new study in the journal Judgment and Decision Making had nearly 1,700 adults read short stories written either by humans or ChatGPT—and the AI tales were consistently rated more engaging and higher quality than the human ones, reports the Guardian. Labels mattered, however: Stories believed to be human-written scored better, regardless of who or what actually wrote them, pointing to a bias toward "authentic" authorship. In follow-up tests, people struggled to tell human and AI fiction apart, often doing no better than chance.

Co-author Deena Skolnick Weisberg says the findings show that AI can now match or exceed human short-story writing, though she argues that doesn't erase the value of human creativity. "AI writing tends to be clearer, more direct, and easier to process," she says, per the BBC. "By contrast, human-written stories are often more subtle and complex. Perhaps people generally prefer predictability, because difficult or subtle material requires more brain power."

Outside critics suggest the findings are overshadowed by AI's ethical and environmental costs. "Of course [AI is] capable of spitting out a coherent and even enjoyable poem with the right prompting and finessing," says University of Birmingham creative writing professor Luke Kennard, who wasn't involved in the research, per the Guardian. "Given that it's been trained on thousands of years of human work, it would be embarrassing if it wasn't." He adds: "I suppose the question we should be asking is: Is it worth it? And the answer is absolutely, categorically, it is not."