Critics warn that AI’s potential for bias, lack of true empathy, and limited human oversight could actually endanger users’ mental health, especially among vulnerable groups like children, teens, people with mental health conditions, and those experiencing suicidal thoughts.

Researchers at Stony Brook University and Columbia Law School found that AI models fine-tuned on a specific author produce texts that readers rate as more stylistically accurate and higher in quality than those written by professional imitators, regardless of the author's publication history.

Readers of texts created to use the styles of famous authors prefer works written by AI to human-written imitations, but only after developers fine-tune AI models to understand an author’s output. This finding, academics argue, means the courts need to rethink assumptions about allowing AI training on authors' works as a fair use exception to copyright liability.

In a preprint paper titled “Readers Prefer Outputs of AI Trained on Copyrighted Books over Expert Human Writers,” Tuhin Chakrabarty, assistant professor of computer science at Stony Brook University, Jane C. Ginsburg, professor of law at Columbia University, and Paramveer Dhillon, associate professor in the School of Information at the University of Michigan, describe how they assessed the impact of AI models that can emulate the style of human writers.