Illustration of Bengalese finches singing (Credit Zachary Jin)
A new study published February 12, 2025, in the open-access journal PLOS Mental Health reveals that responses generated by ChatGPT consistently receive higher ratings than those written by human psychotherapists. The research was conducted by H. Dorian Hatch from The Ohio State University, who is also co-founder of Hatch Data and Mental Health, along with colleagues.
The potential for machines to serve as therapists has gained increasing attention due to various advantages offered by generative artificial intelligence (AI). Previous research has shown that humans often struggle to distinguish between machine-generated and human-written responses. This latest study builds on those findings, demonstrating that AI can produce highly empathic content that is well-received by both mental health professionals and service users – frequently preferred over responses created by human professionals.
Much like how humans arrange words in a particular order to form a grammatically correct sentence, birds tend to sing sets of notes called syllables in a limited number of combinations.
Will you offer us a hand? Every gift, regardless of size, fuels our future.
Your critical contribution enables us to maintain our independence from shareholders or wealthy owners, allowing us to keep up reporting without bias. It means we can continue to make Jewish Business News available to everyone.
You can support us for as little as $1 via PayPal at [email protected].
Thank you.
In their new study involving over 800 participants, Hatch and colleagues showed that, although differences in language patterns were noticed, individuals could rarely identify whether responses were written by ChatGPT or by therapists when presented with 18 couple’s therapy vignettes. This finding echoes Alan Turing’s prediction that humans would be unable to tell the difference between responses written by a machine and those written by a human. In addition, the responses written by ChatGPT were generally rated higher in core psychotherapy guiding principles.
Not only that, just as ChatGPT and other generative language models learn from vast troves of human text to construct flawless sentences, researchers at Penn State have pioneered a groundbreaking method that does the same—but with birdsongs. By training on real avian recordings, their model can generate astonishingly accurate birdsong imitations, unlocking new insights into the intricate structure of these vocalizations and the neurobiology that shapes them.
Beyond revolutionizing the study of birdsong with ChatGPT , this breakthrough could illuminate the very neural mechanisms that govern human language itself, offering a rare glimpse into the deep biological connections between speech and song. The team’s findings, now published in the Journal of Neuroscience, mark a major leap forward in understanding how brains—both avian and human—craft and refine the music of communication.
“Although much simpler, the sequences of a bird’s song syllables are organized in a similar way to human language, so birds provide a good model to explore the neurobiology of language,” said Dezhe Jin, associate professor of physics in the Eberly College of Science and lead author of the paper.
