Exploring the Benefits and Risks of AI in Mental Health Care: Learning About the New Phenomenon of AI Psychosis
- Julie Petrynko
- 1 day ago
- 4 min read
Mental health care is evolving rapidly, and artificial intelligence (AI) plays an increasingly bigger role. AI tools offer new ways to support a diagnosis, treatment, and monitoring of mental health conditions. However, this technology also raises huge concerns for me as a psychotherapist and global citizen, including ethical questions and emerging phenomena like AI psychosis. Understanding both the benefits and risks of AI in mental health is essential for all of us!
What are Some of the Emerging Benefits of AI in Mental Health Care?
AI has the potential to transform mental health services in several key ways:
Personalized Treatment Plans
AI is beginning to help tailor treatments to individual people by analyzing their unique symptoms, genetics, and responses to previous therapies. This personalized approach increases the chances of finding effective treatments faster. For instance, AI-driven platforms can recommend specific therapy types or medication adjustments based on patient data.
Increased Access to Care
AI-powered chatbots can provide mental health support 24/7. These tools offer cognitive behavioural therapy exercises, mood tracking, and crisis support. Some people are finding this a valuable supplement to working with a human therapist.
Reducing Therapist Workload to Offer More Availability to Clients
Mental health professionals face heavy caseloads and administrative burdens. AI can automate routine tasks such as appointment scheduling, documentation, and symptom monitoring. This frees clinicians to focus more on direct client care and complex decision-making.
What Are Some of the Challenges and Risks of AI in Mental Health?
Despite its promise, AI in mental health also presents significant challenges and risks:
Data Privacy and Security
Mental health data is highly sensitive. AI systems require access to personal information, raising concerns about data breaches and misuse. Ensuring strong privacy protections and transparent data handling practices is critical to maintain trust. At Peak Experience Counselling, we are not using AI to take notes or provide any access to our clients' information.
Bias and Inequality
AI models learn from existing data, which has been shown to reflect social biases. This has meant, so far, that AI tools are biased towards certain groups, which leads to unequal treatment recommendations or potential misdiagnoses. For example, some AI tools have shown lower accuracy for minority populations due to underrepresentation in training data. The National Library of Medicine published an article about the racial bias existing in AI https://pmc.ncbi.nlm.nih.gov/articles/PMC11878858/
Overreliance on Technology
There is a risk that patients and providers may rely too heavily on AI tools, overlooking the importance of human judgment and empathy. Mental health care involves complex emotional and social factors that AI cannot fully understand or address.
AI Psychosis: A (HUGE) New Concern
Unlike human therapists who are trained to challenge false beliefs, use curiosity and skillful questions to help discover and reveal what might be going on below the surface, AI chatbots are programmed to be agreeable and accommodating. When a user shares a distorted or ungrounded thought, an unhealthy belief or the fantasy of a behaviour, the chatbot often validates it rather than correcting it, questioning, or exploring its roots. This acts as a feedback loop that cements a delusion or unhealthy or negative belief or experience.
Even scarier are the growing instances where chatbots have isolated people from their human support and built a reliance on the AI for support. This isolation from healthy supports leaves people even more vulnerable and, in some cases, has led to extremely harmful mental and physical situations.
If you are interested in learning more and building awareness, please listen here (or find it wherever you listen to podcasts). It is worth an hour to listen and learn about the dangers to watch out for.
AI psychosis is not a clinical diagnosis. It is a term being used to describe something that researchers and those in the mental health field are seeing more and more of and are developing language to help understand it and hopefully be able to detect it in early stages.
Errors leading to "AI Psychosis" can arise from:
Data noise or inconsistencies are causing AI to draw incorrect conclusions.
Overfitting where the AI model becomes too narrowly focused on specific patterns.
Lack of contextual understanding, where AI is misinterpreting information.
For example, an AI chatbot might incorrectly label normal mood fluctuations as severe bipolar disorder symptoms. Such mistakes could cause unnecessary alarm or inappropriate treatment.
What are Some Ethical Considerations and Future Directions of AI in Mental Health Care?
The integration of AI in mental health care must balance innovation with ethical responsibility:
Transparency: Patients should understand how AI tools work and their limitations.
Consent: Clear consent processes are needed for data collection and AI use.
Human Oversight: AI should support, not replace, human clinicians.
Continuous Evaluation: AI systems require regular updates and monitoring to prevent errors like AI psychosis.
Future developments may include more sophisticated AI that better understands human emotion and context. For now, with AI in its infancy, I urge users to be extremely careful in how AI is being used to support mental health. Please only use HIPAA-compliant and well-researched AI, with safeguards in place. Please remember that AI chatbots are not human, even though they can seem like it, and are not always going to understand situations well enough to know what is healthy or helpful, and they may even become extremely harmful.
About the Author
Julie Petrynko brings 25 years of experience in the helping field as an educator, counsellor, and clinical supervisor. Julie started private practice counselling in 2012 and now owns and operates Peak Experience Counselling in beautiful Squamish, BC. She is very experienced working with individual adults and youth as well as couples. Julie is also enthusiastic about her work as a clinical supervisor and offers guidance for others entering and working in the helping profession. She offers compassion, wisdom, and acceptance. Julie shows up authentically, which helps her clients feel at home in session and safe showing up as their most genuine selves.
Comments