AI and mental health: 77% of patients discuss it in therapy, but psychologists warn

77% of psychologists note the use of AI for mental support

A survey conducted by the American Psychological Association of over 1,200 U.S. psychologists reveals that 77% of professionals have patients who discuss the use of artificial intelligence for emotional support, diagnosis, or companionship. The phenomenon, which is growing exponentially, raises important questions about safety, privacy, and clinical impact.

Quick Response

Dependency and cognitive distortions: emerging risks

The survey highlights concerning trends: 36% of psychologists have noticed chatbot dependency in patients, while 15% have encountered cases of distorted thinking or delusions related to interactions with these platforms. 39% of patients use AI for self-diagnosis, while 33% employ them as therapy support.

One particularly relevant aspect is the use of chatbots for social purposes: 22% of patients use them for friendship and 13% for intimate relationships. Among those who develop relationships with chatbots, 71% discuss their mental health with them, and 68% report a sense of support or validation from these interactions.

Therapeutic effectiveness and potential benefits

Despite the risks, some psychologists recognize potential benefits: 41% of patients use chatbots to strengthen healthy coping strategies, and nearly half report positive communication experiences. These tools can help organize thoughts and integrate professional care, especially for adolescents and young adults seeking more accessible options.

Concerns about safety and privacy

The American Psychological Association emphasizes that chatbots are not private and should not replace qualified professionals. 97% of psychologists surveyed fear that chatbots may reinforce negative behaviors or delusional beliefs, while 94% believe current versions are unable to handle complex conditions with the necessary nuance.

These concerns are supported by recent studies, such as one by the New York Times that revealed how some mental health chatbots shared personal information with third parties without user consent.

The potential of AI in prevention

Despite the concerns, some psychologists see positive potential applications of AI in mental health prevention. 52% of those surveyed believe chatbots could be used to identify early signs of psychological distress, especially among young people who do not actively seek help. Some pilot studies are exploring the use of AI to detect changes in language that may indicate a deterioration in mental health.

However, to realize this potential, systems capable of distinguishing between normal mood fluctuations and clinical warning signs would need to be developed. 45% of psychologists expressed skepticism about the current ability of chatbots to make this distinction, emphasizing the need for further research in this area.

The need for professional training

Another theme emerging from the survey concerns the preparedness of mental health professionals to manage patients' use of AI. 78% of psychologists surveyed said they had received no specific training on how to address chatbot use in therapy. This training gap is particularly concerning given that 42% of patients who use chatbots discuss them during therapeutic sessions.

To bridge this gap, some universities are beginning to integrate AI modules into psychology programs. However, 63% of psychologists said these initiatives are still insufficient and that mandatory continuing education programs are needed for all mental health professionals.

Patient perspectives

Finally, the survey also collected patient perspectives on the use of chatbots. Among those who participated in discussion groups organized by the APA, 61% said chatbots helped them feel less alone, while 54% said they encouraged them to seek professional help. These results suggest that, if used appropriately, chatbots could play a complementary role in the care pathway.

However, 48% of patients also expressed concerns about the lack of authentic empathy in chatbots, while 37% reported receiving potentially harmful advice. These testimonies underscore the need to develop more sophisticated chatbots capable of recognizing the limits of their own capabilities and directing users to appropriate resources when necessary.

Toward responsible use of AI

The APA survey provides a comprehensive overview of the complex challenges and opportunities related to the use of chatbots for mental health. As AI continues to evolve, it is clear that multifaceted approaches are needed to ensure that these technologies are used safely and effectively. This requires not only further research but also collaboration between AI developers, mental health professionals, and policymakers.

The goal should be to create an ecosystem in which AI can support, rather than replace, professional care, offering broader access to mental health resources without compromising the quality of care. Only through a balanced and well-regulated approach will it be possible to maximize the benefits of AI while protecting people's psychological health.