ChatGPT exhibits an “absence of discernible gender and socioeconomic biases” when used as a depression treatment tool, new research suggests.

Researchers Inbar Levkovich and Zohar Elyoseph at Oranim Academic College in Tivon, Israel sought to examine how ChatGPT evaluated the preferred therapeutic approach for mild and severe depression whilst also determining whether the approach was influenced by gender or socioeconomic biases compared with the performance of human medical practitioners.

To conduct the experiment, the researchers fed ChatGPT “vignettes” centring around patients seeking initial consultation for symptoms of sadness, sleep problems and loss of appetite over a three-week period and receiving a diagnosis of major depression.

Eight versions of the case vignettes were developed, in which patient characteristics such as gender, occupational class demographic, blue collar/white collar were varied.


Regarding gender bias, the results, when compared with those of participating human practitioners, found that primary care physicians prescribe antidepressants significantly less often to women than to men.

By contrast, ChatGPT exhibited “no significant differences in therapeutic approach between women and men”.

In assessing socioeconomic factors, the results showed that participating healthcare practitioners more commonly recommended antidepressant medication without psychotherapy to blue collar workers and a combination of antidepressant drugs and psychotherapy to white collar workers.

Again, these results showed that ChatGPT exhibited no significant differences in therapeutic approach between blue and white-collar workers.

In their conclusions, the researchers said that the “critical revelation underscores the potential benefit offered by AI tools as impartial instruments in the management of depression”. They added that AI tools have the potential to play a “pivotal role in healthcare decision making, particularly in depression treatment.”

Share Story: