Top 5 Reason to Be Wary of ChatGPT Results for Finding Treatment
Using an AI search tool to find mental health treatment for a young adult carries major risks that can potentially delay care, lead to unsafe decisions, and waste funds. This is hypocritical coming from me because in 2026 alone, I’ve had four families contact me because I was recommended in their search for treatment. Don’t get me wrong, I’m thrilled that Lilley Consulting is popping up in searches. By my wariness around all of this is that some families are taking the results and not second-guessing the recommendations. So, let’s talk about the top five reasons that anyone searching for anything, especially if it’s healthcare-related in an AI search tool (ChatGPT, Claude, Gemini, etc.).
Privately Owned Algorithmic & SEO Bias
First, AI search engines pull data from scraped web sources, which means their output is heavily influenced by facility search engine optimization (SEO) and marketing budgets rather than clinical outcomes. In the last ten years, I have watched facilities with aggressive online marketing using Google and other search engines. Because of their efforts (and deep pockets), they are likely to be prioritized over a highly regarded, non-profit community program simply because of how data is structured online. Tip: if you are seeing the same company name that pops up for every search, no matter the diagnosis you plug in, be very wary. It’s likely they are not experts in serving all of those populations. You have been warned.
Lack of Medical Vetting and Hallucinated Recommendations
AI models do not verify whether a clinic, therapist, or treatment program is legitimate, accredited, or currently operating. To be direct, they can generate confident-sounding recommendations for facilities that lack proper licensing, have severe safety violations, or do not actually exist. If they don’t exist, that phenomenon is known as AI hallucination. Tip: If you do this search, dig deeper in your own search to identify if the program is licensed, accredited, and if there are any state-violations. Ask a lot of other questions like: Where are your referrals from? Do you pay for leads? Is the program privately owned or a part of a larger behavioral health organization? What is the average number of days clients are getting insurance coverage? What happens when insurance runs out? There’s a lot more you can ask, but these are just some suggestions to give you an idea of what you can (and ought to) consider inquiring about.
Inability to Evaluate Clinical Specificity
Young adult mental health often involves complex, overlapping dynamics such as neurodivergence (ADHD/Autism), trauma, substance self-medication, or mood disorders. If you are only plugging in “young adult treatment,” your results may really off-base. AI cannot conduct a differential diagnosis to determine whether a young adult needs inpatient or residential care, transitional living or intensive outpatient (IOP), or even if you need specialized dual-diagnosis treatment. Relying on an AI search risks selecting a program that is ill-equipped for their actual clinical profile. And if you type in “failure to launch,” God help you. Tip: When you call a program, there is a chance that the Admissions Rep may confirm the bias you believe (in that the program is a match). It’s more realistic that they’re doing their job in “converting a lead” into being a client, when in fact their treatment program is not best for your loved one.
Privacy Vulnerabilities and Absence of HIPAA Protection
This one kills me. Standard consumer AI search engines are not HIPAA-compliant. Read that again. Entering sensitive personal details, psychiatric history, or behavioral descriptions of a young adult into an AI prompt can permanently store that data, potentially using it to train public models or exposing confidential family health information. Unfortunately, I had a parent that I worked with within the last year that copied & pasted her young adult’s neuropsychological evaluation report in an effort to help find programs and resources that aligned with the recommendation. Thinking about this exact scenario makes my stomach flip. Tip: This one is simple - don’t put personal information in an AI search tool. Period. If you’re going to use it to search for options, just let it be the first step. When you call and speak with Admissions professionals about their program, you can also limit the personal information you share with them as well. Don’t share personal information until you know it’s secure.
Zero Real-Time Crisis Assessment or Safety Oversight
Lastly, if a young adult is experiencing acute distress, suicidal ideation, or severe psychosis, AI search tools cannot gauge immediate risk levels or initiate real-world emergency intervention. Typically, parents find themselves flooded emotionally and turn to anyone (or anything) for support in logical and rational guidance. Leaning on AI during a critical transition period can create a false sense of progress while delaying life-saving clinical evaluation though. Tip: Contact your local crisis unit, or take your young adult to the nearest hospital - immediately. It’s never ideal, but it’s better than relying on AI. If you are a part of a parent support community (like REEL, for example), they may have a written “Crisis or Emergency Guide” for you to have for situations like this.
In the end, we’re all searching for something. If you need to find supports for your young adult, please be wary if you are using any AI tool to help narrow down your search. Know that there are also alternatives for finding care. Alternatives such as working with a licensed or credentialed Therapeutic Consultant, or clinical placement specialist who physically audits young adult treatment facilities. Consult a primary care physician, licensed Psychiatrist, or Licensed Therapy for a direct referral. Just make sure that you also ask them when they last toured the program and/or had a client in that facility too.
If you want help finding a program for your young adult contact Joanna.
