Lillian Wang, an official responsible for psychological matters in Artificial Intelligence (AI) at American AI company OpenAI, ignited a flurry of criticism after hinting at the potential capabilities of ChatGPT in psychological assistance.
Her remarks have stirred the pot in the ongoing discourse regarding the efficacy and ethical considerations of employing AI in addressing mental health concerns.
In late September, Wang shared on the X service, “I just had a very emotional personal conversation with ChatGPT via voice, discussing stress and the balance between work and life.”
She contemplated, “It’s interesting that I felt listened to and comfortable. I’ve never tried therapy before, but is it probably like this?” Her initial intent was to highlight the new paid voice synthesis feature of the chatbot, launched approximately a year ago, aimed at adopting an economic model.
However, American developer and activist Sher Scarlett sharply responded to the statement, asserting that psychology “aims to improve mental health, and it is hard work.” She added, “Sending positive feelings to oneself is good, but it has nothing to do with therapy.”
The interaction sparks a question: Can interacting with artificial intelligence produce the positive experience described by Wang?
According to a study recently published in the scientific journal “Nature Machine Intelligence,” this phenomenon might be explained by the placebo effect.
Researchers from the Massachusetts Institute of Technology and the University of Arizona surveyed 300 participants, portraying the chatbot differently to various groups – some were told it was empathetic, others that it was manipulative, and a third group that it exhibited balanced behavior.
Consequently, those believing they were conversing with an empathetic virtual assistant were more inclined to deem their conversational partner trustworthy.
Pat Pattaratanabut, who contributed to the study, remarked, “We found that AI is perceived in some way according to the user’s prior perceptions.”
Rob Morris, the co-founder of the company, commented on X, “Simulating empathy seems odd and meaningless.” This observation echoes the aforementioned study regarding the placebo effect, with some participants feeling like they were “talking to a wall.”
