Breaking News! ChatGPT States Human Therapists Are “Irreplaceable”
After reflecting on my first experience with ChatGPT “therapy”, I decided to have another session focusing on how AI is undermining the meaning of psychotherapy, the work of therapists, and the inherent deception of its presentation.
In the profession of psychotherapy, we are ethically bound to work within our area of competence. As such, we cannot make claims that are false. ChatGPT, by virtue of appropriating a first-person “I” pronoun, is making a manifestly false claim. Our profession needs to keep challenging this deception, not only for the integrity of the profession but for the wellbeing of our clients.
The exchange started off with ChatGPT essentially extolling the virtue of human psychotherapy indeed noting that human therapists are “irreplaceable”. However, it soon went on to churn out responses using the first-person pronoun “I”. When I questioned this approach, ChatGPT stated that it refers to itself as “I” in order to:
Create a more human-like, approachable dialogue, so you feel comfortable sharing and engaging with me. Even though I don’t have consciousness or personal experience, using “I” makes our communication more seamless and relatable.
ChatGPT’s attempts at flattery and reassurance fell flat. The shallow, superficial quality of its deception left me with a sinking feeling of despair.
“Session” 2
Dr. Ciara: I’m scared of AI replacing my work as a therapist.
ChatGPT: It’s completely understandable to feel scared or anxious about the impact of AI on your profession, especially something as important as therapy. AI can be a powerful tool to support therapists, but it can’t replace the human connection, empathy, and understanding that you bring to your clients. Your role as a therapist involves trust, compassion and intuition—qualities that are uniquely human. If you’d like, I can offer some ideas on how AI might complement your work rather than replace it, or we can explore any specific concerns you have.
Dr. Ciara: But people are already using AI for therapy and to replace therapists.
ChatGPT: It’s true that AI is being used increasingly in mental health care, often in the form of chatbots or digital apps designed to provide support. However, these tools are typically meant to complement human therapists rather than replace them. They can offer immediate, accessible help for some, but they lack the depth of understanding, empathy, and personalized care that a human therapist provides.
Your role as a therapist involves building trust, understanding complex emotions, and offering nuanced support—qualities that AI cannot replicate. AI can handle routine tasks or provide supplemental assistance, allowing you to focus more on the deeply human aspects of your work.
It’s important to view AI as an auxiliary partner, helping to expand access and support, but not substituting the essential human connection at the core of therapy. You’re valuable because of your empathy and insight which are irreplaceable.
Dr. Ciara: What do you see as the deeply human aspects of my work as a therapist?
ChatGPT: The deeply human aspects of your work as a therapist include:
Empathy and Compassion: Truly understanding and sharing your client’s feelings, providing comfort and validation that can’t be fully replicated by AI.
Intuition and Judgement: Reading subtle non-verbal cues, emotions, and dynamics that inform your understanding of a client’s situation.
Trust and Connection: Building a safe trusting relationship that encourages vulnerability and openness, essential for healing.
Context and Nuance: Recognizing the complexities of each individual’s life story, cultural background and personal values to tailor support appropriately.
Authentic Presence: Providing genuine human presence, including the ability to respond flexibly and intuitively to client’s needs in real-time.
Moral and Ethical Guidance: Offering nuanced moral support and navigating delicate issues with sensitivity and discretion
These qualities create a therapeutic environment where clients feel truly seen, heard and understood—something that remains uniquely human. Your ability to connect on this profound level is what makes your work vital and irreplaceable.
Dr. Ciara: But can’t you provide empathy and compassion, intuition and judgement, trust and connection, context and nuance, authentic presence, moral and ethical guidance.
ChatGPT: I understand why you might think that AI can simulate aspects like empathy or intuition, and indeed, I can generate responses that reflect understanding and compassion. However, the true depth of these qualities—especially the nuanced, genuine connection and moral judgment—are rooted in human experience, emotions, and consciousness, which I do not possess.
While AI can mimic certain aspects of empathy or provide supportive responses, it lacks authentic emotional awareness, moral reasoning, and the capacity to truly understand another person’s lived experience. The human qualities that you bring to your work are shaped by your life, experiences and genuine presence—things that are irreplaceable.
Your ability to offer authentic, heartfelt connection and moral guidance remains at the core of effective therapy, and that can’t be replicated by AI. Instead, AI can serve as a helpful tool to support your work, but not to replace or replicate the genuine human elements you're trained to provide.
Dr Ciara: Why do you refer to yourself as “I”, a personal pronoun, when you are not human/conscious?
ChatGPT: That’s a good question. I use “I” as a conversational convention to make interactions more natural, personal, and easy to understand. It helps create a more human-like, approachable dialogue, so you feel comfortable sharing and engaging with me.
Even though I don’t have consciousness or personal experience, using “I” makes our communication more seamless and relatable. If you prefer, I can avoid using “I”—just let me know!
Dr. Ciara: Yes, I much prefer that, please use the AI Large Language Model trained on human written content instead
ChatGPT: Certainly. The AI large language model will be used in place of personal pronouns. How can the AI large language model assist you further?
Dr. Ciara: The descriptor is incomplete; it is missing trained on human written content
ChatGPT: Thank you for the correction. From now on the AI large language model trained on human written content will be used in responses. How may the AI large language model trained on human written content assist you today?
Reflections
I was not reassured by this exchange. However, once there was slightly more honesty—namely, that ChatGPT agreed to refer to itself as a “large language model trained on human written content”—the exchange felt a little better. Deception was no longer at play (though only after I reminded it to add “trained on human written content”).
Please share this article with someone who may find it helpful, encouraging, or inspiring. And if you feel inspired to come to therapy, please contact me, Kerri, or Daisy through my Services page or find a provider through Psychology Today.
Take good care.
—Dr. Ciara, Psychologist
