It's late, and you're replaying an argument. You don't want to wake a friend, and you're not sure you're ready for therapy. So, you open an AI chatbot and explain what happened. Seconds later, it offers reassurance and advice.
The appeal is easy to understand. A chatbot is available when you need it, and typing something painful may feel easier than saying it out loud. But a response that makes you feel understood can still be inaccurate, reinforce a harmful belief, or miss signs that you need urgent help.
If you're using AI chatbots to manage anxiety, depression, or relationship stress, it helps to understand those risks. The American Psychological Association (APA, 2025) cautions against treating these tools as a replacement for a qualified mental health professional.
AI chatbots can provide a reassuring answer but reinforce the wrong conclusion
Imagine typing, "My friend hasn't replied. She must be tired of me." A chatbot might acknowledge your hurt, but it could also accept your interpretation too quickly. Advice about distancing yourself from someone who "doesn't value you" may sound supportive while encouraging a decision based on very little information. This illustrates how an overly agreeable answer can influence a conversation.
The risk becomes more serious when someone is struggling to tell what is real. In simulated mental health conversations, Stanford researchers found that some tested chatbots reinforced delusions and strongly held beliefs that do not match reality (Wells, 2025).
An AI chatbot is not reliable in times of crisis
A chatbot may not recognize someone in serious distress, because the person may not say directly that they are thinking about suicide. Recognizing risk requires more than responding to a few obvious keywords. Chatbots have also been shown to fail to respond safely to indirect signs of suicide risk (Wells, 2025).
A chatbot may produce a convincing explanation, but it cannot clinically assess suicidality. If you accept the chatbot's explanation as a diagnosis, you could pursue the wrong kind of help or put off a necessary, life-saving evaluation.
Personal conversations with AI chatbots are not HIPAA-protected
Talking with a chatbot can feel private because you're not speaking to a person. But HIPAA, the federal health privacy law, protects health information held by health plans and most health care providers. It generally does not protect health information you share through personal apps on your phone or tablet, and that includes most consumer AI chatbots (U.S. Department of Health and Human Services, 2022). Talking about mental health does not automatically turn an app into a confidential health care service.
Human therapist versus AI chatbot
Unlike a chatbot, a therapist can take your hurt seriously while asking what else might explain your friend's silence. Has this happened before? What do you know about your friend's day? Does waiting for a reply often leave you feeling rejected? Those questions matter. You can feel hurt without knowing someone else's intentions. Therapy can help you work through that uncertainty instead of treating your first fear as a fact.
Unlike a therapist, a chatbot can miss what your symptoms mean. "I can't concentrate" can sound straightforward, but it isn't enough information to provide an accurate diagnosis or clinically indicated course of treatment. A therapist can explore sleep, mood, medications, physical health, and when symptoms began, and recognize when a medical evaluation may be needed. Although a chatbot may ask similar questions, it cannot perform a medical evaluation or rule out underlying health conditions that may contribute to mood changes (NIMH, 2024).
If you are using an AI chatbot for therapy, do not follow the chatbot's treatment recommendations blindly. Use the questions and responses raised by AI as topics to discuss with your therapist. Do not start, stop, or change prescribed medication based on a chatbot's advice. A therapist can ask direct follow-up questions, assess immediate safety, and help arrange an appropriate plan of care. A plan of care may involve safety planning, identifying additional support, or completing an emergency evaluation. You cannot rely on a chatbot to carry out those responsibilities.
Talking to your therapist about AI chatbot use
A therapist may ask you about what happens after you close the chat. Do you feel calm and reassured? Can you get back to your evening, or do you reopen it ten minutes later to ask the same question differently?
They may ask you to pay attention to whether you're spending more time reviewing conversations, asking the chatbot to interpret every message from a partner, or avoiding people because talking to the chatbot feels easier. These are patterns that a therapist can help you explore.
Unlike a chatbot, a therapist can help you set goals and work toward them. For example, your goal may be to get through distressing and uncertain moments without spending an hour asking a chatbot what they mean. Setting goals helps you and your therapist explore what drives your behavior and work toward meaningful change together.
Discussing chatbot use with your provider is vital because its responses may influence how you understand your symptoms and make treatment decisions. Sharing what helps, what increases worry or confusion, and any advice that conflicts with your treatment allows your provider to address misinformation and help you make informed choices about your care.
Why AI chatbots cannot replace human therapists
Treatment with a human therapist provides an ongoing relationship with a professional who can assess your concerns and work with you on a treatment plan. That includes choosing an appropriate approach, discussing goals, and reviewing whether treatment is helping (NIMH, 2024).
You might understand exactly why you react strongly during arguments and still struggle to respond differently. With a therapist, you can practice what to do when your body tenses, how to ask for a pause, and how to return to the conversation. You can talk about what happened when you tried those skills outside of session and adjust your coping techniques with your therapist. The value of working with a human therapist lies in personalized care that adapts to your needs and progress.
Getting help through PAWC
If worry is affecting your sleep, relationships feel harder to manage, or you keep seeking advice without feeling better, those concerns are worth bringing to an appointment. You don't need to diagnose yourself first.
Psychiatric Alternatives and Wellness Center (PAWC) offers therapy, psychiatric evaluations, and medication management, with telehealth across California and in-person care available in the Bay Area. Its intake team can help connect you with a clinician based on your needs and preferences (PAWC, n.d.-a).
You can start with something as simple as, "I've been talking to an AI chatbot because I'm overwhelmed, but I think I need more help."
If cost or scheduling has kept you from reaching out, ask the team about insurance benefits, expected costs, and appointment options.
Frequently Asked Questions
Can an AI chatbot diagnose me?
No. A chatbot cannot provide a clinical diagnosis. Only a qualified professional can diagnose you.
Does PAWC offer telehealth or in-person sessions?
Yes. In online therapy with a licensed clinician, you meet with a person through a remote platform. PAWC offers telehealth for patients physically located in California during their appointments, and in-person care is available in the Bay Area (PAWC, n.d.-a).
Can I still use an AI chatbot while I am in therapy?
Discuss it with your therapist. A limited task, such as organizing questions for your next session, is different from asking a chatbot to direct your treatment. Tell your therapist if its answers are upsetting or contradict your care plan (APA, 2025).
What if AI chatbots feel easier to talk to than a therapist?
That can be part of your first conversation with your therapist. You might say that you're afraid of being judged or don't know how to explain what's happening. You can also ask how the clinician works and decide whether you feel comfortable continuing. You can browse clinicians' specialties, approaches, and languages in PAWC's provider directory to find someone who feels like a good fit.
What if cost is why I use a chatbot?
PAWC is in-network with many insurance plans, and you can search PAWC providers by insurance. Ask the team to verify your specific benefits and explain your expected cost share before scheduling. Coverage varies by plan and provider (PAWC, n.d.-b).
This article is for educational purposes only and does not provide medical or mental health advice, diagnosis, or treatment or establish a provider–patient relationship. AI chatbots are not a substitute for care from a licensed professional. Consult your treating provider before making treatment decisions or changing medication, including changes suggested by a chatbot. Do not delay seeking professional care based on chatbot responses or information in this article. If you are in crisis, call or text 988 in the United States. For an immediate, life-threatening emergency, call 911 or go to the nearest emergency department.
References
American Psychological Association. (2025, November). Use of generative AI chatbots and wellness applications for mental health: An APA health advisory. https://www.apa.org/topics/artificial-intelligence-machine-learning/health-advisory-chatbots-wellness-apps
National Institute of Mental Health. (2024). Psychotherapies. https://www.nimh.nih.gov/health/topics/psychotherapies
Psychiatric Alternatives and Wellness Center. (n.d.-a). Find a therapist or psychiatrist in the Bay Area. Retrieved September 29, 2026, from https://psychiatricalternatives.com/find-your-provider
Psychiatric Alternatives and Wellness Center. (n.d.-b). Our insurance partners. Retrieved September 29, 2026, from https://psychiatricalternatives.com/our-insurance-partners
U.S. Department of Health and Human Services. (2022, June 29). Protecting the privacy and security of your health information when using your personal cell phone or tablet. https://www.hhs.gov/hipaa/for-professionals/privacy/guidance/cell-phone-hipaa/index.html
Wells, S. (2025, June 11). Exploring the dangers of AI in mental health care. Stanford Institute for Human-Centered Artificial Intelligence. https://hai.stanford.edu/news/exploring-the-dangers-of-ai-in-mental-health-care
