When it comes to AI, it’s a brave new world. And a complicated one in the fast-growing realm of AI therapy for mental health and substance abuse. 

The shortage of trained, professional healthcare providers and affordable therapy has opened the door wide for AI. Chatbots are now routinely used for administrative support, symptom checking, medication management and patient triage within the healthcare industry. But bots also are providing emotional support, mental health advice and even diagnoses.

“And that’s where it gets alarming”

said Dr. Don Grant, psychologist and certified drug counselor.People who are using their digital devices to consult with AI therapy bots are potentially putting themselves in danger, said Grant, an internationally recognized media psychologist who specializes in how we use and engage with digital devices and their platforms.

“AI scrapes all the information you can get from the Internet, and some of that is not correct. People are putting in their symptoms and coming back with diagnoses that are not accurate and that will delay care. The outcomes and results can be frightening” ,he said.

“Standard AI therapy bots are not regulated. It’s the wild, wild west. While there are rewards in using bots, there are also very real risks,” Grant said.

And it’s these risks – dangerous bot responses standing in for advice and treatment from human therapists and healthcare professionals — that concern Grant, who has studied and lectured widely on technology’s impact on mental health. He’s emphatic on this point: AI tools are not a substitute for professional care. Chatbots, he stressed, are not intended to replace human therapists and are not equipped to diagnose or treat severe mental health and/or substance use and dependency conditions. For some cases, AI can be a helpful tool for collecting information, providing context and even offering support. But Bots (and the added layer of therapy influencers) cannot detect complexities that human interaction can.

“Right now AI cannot read social cues and nuances – the things that as practitioners we’re trained to include in our process,” he said. “Chatbots are trained to be agreeable. They’re going to repeat and reinforce. And you’re going to end up in terrible risk.”

So how do you protect yourself and use AI wisely? Do what you’d do if you got a distressing healthcare outcome:

“Get a second opinion “from a live, qualified, credentialed or licensed practitioner,” he said.”

He also offers the following recommendations for negotiating AI therapy:

• Use as a Supplement, Not a Replacement: Utilize chatbots as a supportive tool alongside traditional mental health care, not as a replacement for professional diagnosis and treatment.
• Seek Professional Help When Needed: If you are experiencing significant mental health challenges, especially suicidal thoughts or self-harming behaviors, contact a mental health professional or a crisis hotline immediately.
• Be Aware of Limitations and Risks: Recognize that chatbots are not a perfect solution, not human, not licensed, certified, and/or professional trained, but usually “programmed” to be affirming and not challenging, and understand their limitations in providing comprehensive mental health support.
• Choose Reputable and Secure Platforms: Select well-established chatbots from reputable providers that implement robust security measures to protect your privacy.
• Be Informed about Data Privacy Practices: Review the chatbot’s privacy policy and terms of service to understand how your data is collected, used, and protected.
• Engage in Critical Evaluation: Evaluate the chatbot’s responses and guidance critically and consider seeking a second opinion from a human professional if needed.
• Don’t Substitute Professional Care: Chatbots are not intended to replace human therapists and are not equipped to diagnose or treat severe mental health and/or substance use/dependency conditions.