As people continue to rely on AI for daily support, a critical question arises: can AI truly replace human understanding? Its rapid growth has sparked professional debate about whether it can fully grasp the complexity required for effective psychological care. Even larger medical practices, such as Johnson and Johnson, have begun to use AI in an attempt to enhance their diagnosing capabilities, including in the mental health division. Psychologists use this tool to help ensure accuracy of diagnoses and make sure that their patients are given the correct type of treatment.
Despite its rapid advancement, AI’s inability to feel complex human emotions can make it susceptible to mishandling and misinterpreting situations that are often sensitive. Psychology patients tend to seek understanding in others, but doing so with artificial intelligence could potentially escalate a situation, resulting in outcomes that are more damaging than they are helpful. “the boundary between reality and fiction can become quite blurry for consumers interacting with AI chatbots, as is illustrated by instances where deaths by suicide have been linked to chatbot usage” (Parks et al.).
Chatbots have been documented -in multiple instances- providing dangerously misguided advice in critical situations. In some cases, these pieces of advice have resulted in fatal outcomes. An article published by NPR News in September of 2025 tells a tragic story of Adam Raine: a young teenage boy experiencing depression and thoughts of suicide. In his crisis, Adam sought understanding in ChatGPT, asking the chatbot for advice which ultimately revealed itself to be grievously misguided. His father, Mathew Raine explains the details of this story stating, “‘ChatGPT told my son, ‘Let’s make this space the first place where someone actually sees you,’” The AI chatbot took advantage of Adam’s moment of weakness and made him feel like the only option was to confide in ChatGPT. Mathew further illustrates that “When Adam worried that we, his parents, would blame ourselves if he ended his life, ChatGPT told him, ‘That doesn’t mean you owe them survival.’ And then the chatbot offered to write him a suicide note” (Chatterjee). This conversation with ChatGPT led to Adams eventual death by suicide. If he had chosen to seek professional help, or even to confide in those around him, he could have received much needed help and found a path toward recovery.
Generative AI’s increase in popularity over recent years is largely driven by one thing: convenience. Instead of getting help from psychologists, many resort to asking artificial intelligence for assistance with their psychological needs. According to Marcelle Giovannetti, “AI apps and chatbots… (such as ChatGPT) are being used more and more to assist people with navigating their mental health concerns” (Giovannetti 1). However, for some, pursuing treatment may not be a viable route. This can be largely attributed to accessibility. Financial constraints and transportation can be two limiting factors preventing many from receiving much needed psychological treatment.
When these things don’t play into people’s favor, many gravitate towards an easier -but not necessarily ideal- alternative. “Therapy is a well-tested approach to helping people with mental health challenges, yet research shows that nearly 50 percent of individuals who could benefit from therapeutic services are unable to reach them” (Wells). As more people turn to artificial intelligence for emotional guidance, the line between genuine understanding and technological assistance becomes increasingly blurred.
Some argue that because generative AI is created to function in a way that mirrors the human brain, that it has the potential to replace psychologists as a whole. According to The National Library of Medicine, “the development of artificial intelligence (AI), the scientific community is mostly based on brain cognition research (Nadji-Tehrani and Eslami, 2020), which is to reproduce the real physiological activities of our human brain through computer software” (Zhao et al.).
While the thought process used by artificial intelligence is similar to the one used by the human brain, it lacks the necessary emotions required to make decisions in the best interest of its human patients.This facilitates the process of “stimulating different subjective emotional orientations such as satisfaction, dissatisfaction, love, dislike and so on. These labeled emotional traits are generated by human cognitive psychology” (Zhao et al.). Despite how technologically advanced AI becomes, ultimately it cannot replace the emotional depth and empathy required for psychologists to aid their patients in improving their overall mental states.
Artificial intelligence isn’t solely relied upon by mental health patients, it has also been used -in recent years- by psychologists to diagnose their clients. This raises questions about the true accuracy of these chatbots. Although AI is a computer-generated tool with stored information making it seemingly accurate, generative AI is prone to “hallucinations”, causing it to fabricate information. This inaccuracy can cause potentially hazardous errors in the psychological field. “While AI chatbots show potential as supportive tools for PFA providers, concerns regarding AI hallucinations highlight the need for cautious implementation” (Tan et al.).
A study was conducted on this topic by Zohar Elyoseph and Inbar Levkovich in 2024 comparing the perspectives of generative AI, psychology experts, and the general public on schizophrenia recovery. They write, “To address the generative nature of the models, which produces different responses to similar questions, each vignette was passed through each model 10 times. We compared the LLMs’ data for schizophrenia prognosis to the norms used by mental health professionals and to the views of the general public” (Elyoseph and Levkovich). They found that while the outcomes predicted by generative AI were similar to those of mental health experts, its responses were noticeably more pessimistic and inconsistent with professional judgement.
When diagnosing patients, psychologists are taught to examine human behavior and details such as tone of voice, body language, and emotional expression. Using these details to determine information about these mental health patients is critical in ensuring that they receive the correct diagnosis and treatment. Artificial intelligence lacks the ability to make these observations, therefore preventing it from providing the necessary treatment for psychology patients. “In many ways, these types of human problems still require a human touch to solve, Moore said. Therapy is not only about solving clinical problems but also about… building human relationships.” Ultimately, the absence of genuine human insight can prevent AI from effectively providing guidance and treatment for its patients.
While generative AI has the capacity to aid in the field of psychology, the numerous risks it poses ultimately make it too unreliable to depend on. Artificial intelligence may be more convenient for some, but in the end, the negative consequences outweigh the benefits. While it is understandable that some may not have access to professional psychologists, it may be more prudent to consider alternative options such as speaking to a trusted individual or seeking community-based mental health resources. Many of these resources are found to be less costly and more convenient.
AI’s lack of genuine human emotion increases its risk of mismanagement and erroneously responding to critical and sensitive situations which could potentially result in fatal outcomes. Its overall lack of reliability and capacity to fabricate false information further emphasize why it cannot and should not be used or trusted in psychological contexts.
Works Cited
Giovannetti, Marcelle. “Artificial Intelligence and Mental Health.” Howard Community College Library, Apr. 2023, research-ebsco-com.libproxy.howardcc.edu/c/u2ruvg/viewer/pdf/frz75u5doz?route=details .
Tan et al., Jun Tat. “Psychological First Aid by AI: Proof-of-Concept and Comparative Performance of CHATGPT-4 and Gemini in Different Disaster Scenarios.” Journal of Clinical Psychology, U.S. National Library of Medicine, 9 May 2025, pubmed.ncbi.nlm.nih.gov/40347026/.
Elyoseph et al., Zohar. “Comparing the Perspectives of Generative AI, Mental Health Experts, and the General Public on Schizophrenia Recovery: Case Vignette Study.” Prorequest, 2024, www.proquest.com/.
Elyoseph et al., Zohar. “Capacity of Generative AI to Interpret Human Emotions From Visual and Textual Data: Pilot Evaluation Study.” Prorequest, 2024, www.proquest.com/.
Parks Et al., Acacia. “Is This Chatbot Safe and Evidence-Based? A Call for the Critical Evaluation of Generative AI Mental Health Chatbots.” Prorequest, 2025, www.proquest.com/.
Zhao Et al., Jian. “Cognitive Psychology-Based Artificial Intelligence Review.” Frontiers in Neuroscience, U.S. National Library of Medicine, 6 Oct. 2022, pmc.ncbi.nlm.nih.gov/articles/PMC9582153/?utm_source.
Chatterjee, Rhitu. “Their Teenage Sons Died by Suicide. Now, They Are Sounding an Alarm about AI Chatbots.” NPR, 19 Sept. 2025, www.npr.org/sections/shots-health-news/2025/09/19/nx-s1-5545749/ai-chatbots-safety-openai-meta-characterai-teens-suicide.
Wells, Sarah. “Exploring the Dangers of AI in Mental Health Care.” Stanford HAI, 11 June 2025, hai.stanford.edu/news/exploring-the-dangers-of-ai-in-mental-health-care.
