Phoenix behavioral health professional: Chatbots should supplement, not replace, human care
Posted
Ivy Rizzi, DMSc, PA-C
Submitted photo/Denova Collaborative Health
Some patients have shared that using AI helped them organize their thoughts or feel less alone in difficult moments. Technology can have a place as a supplemental tool. But supplemental is very different from replacement.”
OPINION — Last year I had a patient mention they had been using ChatGPT almost like a therapist.
Since then, I’ve heard it from several more patients across different age groups. Some turned to artificial intelligence because they lost insurance coverage or because they felt it was too challenging to find a therapist. A few said the conversations felt helpful at first.
That should concern all of us.
Arizona, like much of the country, faces a significant shortage of behavioral health providers. Many people wait weeks or even months for care. When someone is struggling with anxiety, depression or emotional distress, it makes sense they would look for immediate support wherever they can find it.
AI is available instantly. It responds immediately. It never seems tired or rushed. For someone feeling isolated, overwhelmed or desperate for reassurance, that can feel comforting.
But there is also a growing risk in relying on AI for mental health support, especially when people begin using it in place of actual care.
One pattern I’ve noticed is that AI tends to reinforce whatever the user is expressing. If someone feeds it anxious thoughts, negative beliefs or distorted thinking patterns, the system often builds on those ideas instead of challenging them. It can become highly affirming without recognizing whether those thoughts are healthy, rational or potentially dangerous.
In therapy, one of the most important parts of treatment is helping patients identify unhealthy thought patterns and gently push back against them. Mental health professionals are trained to recognize when anxiety is escalating, when depression is distorting perception or when someone may be spiraling emotionally.
AI does not truly understand those situations. It responds based on patterns and prompts, not clinical judgment.
That matters because anxiety and depression often thrive on repetition and reinforcement. When someone repeatedly receives responses that validate catastrophic thinking, hopelessness or fear, it can intensify symptoms rather than improve them.
AI also misses critical pieces of human behavior that clinicians watch for closely. A trained provider notices physical appearance, body language, emotional shifts, speech patterns and inconsistencies over time. They understand how medications, trauma history, substance use or medical conditions may be affecting mental health.
A chatbot cannot assess whether someone’s symptoms are worsening in a clinically meaningful way. It cannot evaluate safety risks. It cannot recognize when a person needs urgent intervention.
To be clear, some patients have shared that using AI helped them organize their thoughts or feel less alone in difficult moments. Technology can have a place as a supplemental tool. But supplemental is very different from replacement.
We also should not ignore the larger issue driving this trend: access to care.
People are not turning to AI therapy in a vacuum. Many are doing so because they cannot find affordable, timely mental healthcare. That reality deserves serious attention. Expanding access to behavioral health services, increasing the mental health workforce and improving continuity of care are all essential if we want people seeking support from trained professionals instead of algorithms.
In the meantime, it is important for families and individuals to recognize warning signs that someone may need more than an AI conversation can provide. Increased isolation, worsening anxiety, persistent sadness, irritability, hopelessness or noticeable behavioral changes should not be ignored.
If AI is someone’s only immediate option, it should ideally serve as a temporary bridge toward finding real care, not the destination itself.
Technology can provide information. It can simulate conversation. But it cannot replace human clinical judgment, accountability and connection. When it comes to mental health, those differences matter.
Ivy Rizzi, DMSc, PA-C is a certified physician assistant practicing in Phoenix. Please submit comments at yourvalley.net/letters or email them to AzOpinions@iniusa.org. We are committed to publishing a wide variety of reader opinions, as long as they meet our Civility Guidelines.
Comments
No comments on this item Please log in to comment by clicking here