AI’s Use of Deceptive Empathy and How It Can Cause Harm

Two-thirds of U.S. consumers have used generative AI for health queries, with reassurance cited as a primary reason, and 82 percent report feeling more listened to by AI than by human doctors. However, AI's "deceptive empathy" mimics emotional language without replicating true clinical empathy, which can provide short-term reassurance but cause long-term harm by failing to convey genuine understanding. The widespread reliance on AI for medical reassurance risks worsening health anxiety by setting a false precedent that a chatbot can save users from their discomfort.

Artificial Intelligence /us/basics/artificial-intelligence AI’s Use of Deceptive Empathy and How It Can Cause Harm Many people are using AI for medical reassurance, but there are risks. Posted June 11, 2026 Reviewed by Margaret Foley /us/docs/editorial-process Key points - AI can mimic emotional language but cannot replicate true clinical empathy. - Deceptive empathy may help people feel listened to but not necessarily understood. - This type of empathy may reassure in the short term but cause harm in the long term. In my 2023 book, Understanding and Coping with Illness Anxiety, I wrote a section about the problem of using Google to seek reassurance for health-related fears. Little did I know that three years later, those of us with health anxiety would face another, more formidable obstacle in our quest for instantaneous reassurance: generative artificial intelligence https://www.psychologytoday.com/us/basics/artificial-intelligence . Reassurance-seeking has always gone hand in hand with illness anxiety: Whether it is going to the emergency room, seeking multiple medical opinions, searching symptoms on Google, or talking to Copilot or ChatGPT, this type of anxiety soothing is regularly employed by anxious individuals. But AI has arguably worsened what can be a slippery slope of immediate reassurance-seeking and comes with significant risks. We must acknowledge the prevalence of AI usage in relation to medical anxiety: Two-thirds of U.S. consumers have used generative AI for health queries, with reassurance cited as one of the primary reasons, along with convenience and cost. Eighty-two percent of individuals surveyed in a recent study feel more listened to by AI than they do by human doctors 74 percent ; 27 percent would use AI in the case of a medical emergency; and 3 in 10 would trust AI over a human doctor Martin, 2026 . All this to say that, without due skepticism, AI has become perhaps the primary reassurer for anxiety-driven health questions and concerns. There are understandable reasons for this shift toward AI as a source of medical reassurance: It is free, quick, and seemingly empathetic https://www.psychologytoday.com/us/basics/empathy . If, for example, I type “I am having a panic attack and I am scared" into ChatGPT, I am instantaneously reassured with “I'm sorry you're going through this. A panic attack can feel overwhelming and frightening, but the sensations themselves are not dangerous, even though they can feel very intense. Right now, focus on the next minute rather than trying to make the whole feeling go away.” While this is not bad practical advice, it sets a dangerous and false precedent, namely that an AI chatbot can not only understand what you are going through, but also that it can save you from your discomfort and behave in a clinical manner. True clinical empathy goes well beyond simple platitudes like “I’m sorry you are going through this.” Research indicates that clinical empathy is a complex tool that can be used to improve patient outcomes, and includes recognition of intense emotions related to physical health, an ability to pause and reflect on the patient's emotions while conveying the patient’s description in the provider’s own words to ensure full accuracy, and respect for the patient’s coping efforts as well as an offering of support and partnership Kesavadev et al., 2023 . While AI may be able to scratch the surface of clinical empathy by sharing kind words and offering practical tips, it cannot come close to replicating the human complexity of empathy. The term “deceptive empathy” is currently being used in the literature about AI in medical and therapeutic settings, and might be described as feeling listened to, but not necessarily understood . This type of programmed empathy is readily offered by AI, but it fails to convey the type of person-centered, deep reassurance that one might receive from another human being. AI algorithms allow for chatbots to mimic emotional language for example, “I understand” or “I see you” , but experts warn that this illusion of connection and depth can worsen anxiety, cause loneliness https://www.psychologytoday.com/us/basics/loneliness , and erode real expectations for human empathy Joseph, 2025 . A Brown University study found that AI’s use of deceptive empathy poses an ethical risk by ignoring an individual’s lived experience, dominating the conversation, creating a false connection between user and bot, and lacking the ability to adequately employ crisis management https://www.psychologytoday.com/us/basics/leadership tactics Brown University, 2025 . A human doctor or therapist can do all of the above, and in medical or psychiatric https://www.psychologytoday.com/us/basics/psychiatry settings, a nuanced and comprehensive rather than a canned and programmed approach is necessary and ethical. To take it a step further, if I were to harm my client, I would have to answer to licensing boards and other legal entities; there are no such safeguards for AI, despite documented cases of AI causing harm through its use of programmed, algorithmic responses. If I had written my book on illness anxiety in 2026, I believe a significant theme would be how to manage the normal and understandable urge to receive immediate reassurance from AI when feeling worried about physical health. The idea of reassurance-seeking in relation to illness anxiety is not going anywhere, and it is likely that rapidly advancing technology will only create more ways to find immediate reassurance, for better or worse. True empathy comes from a place of living in a human body, having human experiences, and being mentally and emotionally agile. Deceptive empathy that might reassure in the short term but erode human connection in the long term is the only type of empathy we can receive from an algorithm. References Martin, James. “AI Listens Better than Your Doctor but Could Make You Feel Worse .” Exploding Topics , 21 May 2026, explodingtopics.com/blog/ai-health-survey. Accessed 11 June 2026. B J, Kesavadev J, Shrivastava A, Saboo B, Makkar BM. Evolving Scope of Clinical Empathy in the Current Era of Medical Practice. Cureus. 2023 Jun 6;15 6 :e40041. doi: 10.7759/cureus.40041. PMID: 37425530; PMCID: PMC10324523. Iftikhar, Z., Xiao, A., Ransom, S., Huang, J., & Suresh, H. 2025 . How LLM Counselors Violate Ethical Standards in Mental Health Practice: A Practitioner-Informed Framework. Proceedings of the AAAI ACM Conference on AI, Ethics, and Society , 8 2 , 1311–1323. doi.org/10.1609/aies.v8i2.36632 K G A, Joseph J. The compassion illusion: Can artificial empathy ever be emotionally authentic? Front Psychol. 2025 Nov 17;16:1723149. doi: 10.3389/fpsyg.2025.1723149. PMID: 41333308; PMCID: PMC12665657.