The U.S. is in the midst of an acute mental health crisis . In late May, the Centers for Disease Control and Prevention reported that 19% of adults have been told by a physician or therapist that they have a depressive disorder and 29% of U.S. high school students report that their mental health is precarious. While the causes are myriad — inflation, blatant authoritarianism, breathtaking wealth inequality — the artificial intelligence boom is destabilizing systems that were already close to their breaking point. In the realm of mental health care, changes in health insurance coverage for long-term psychotherapy, coupled with the emergence of AI chatbots that masquerade as counselors, have put access to high-quality, affordable care in serious jeopardy.
“The use of technology is way ahead of evaluation,” Linda Michaels, chair and co-founder of the Psychotherapy Action Network (PSiaN), tells Truthdig. “We know nothing about the long-term efficacy of AI therapy. Vulnerable people who are dealing with abuse, trauma, addiction or personality issues need to process what happened to them. But some of the AI platforms have announced a single-session tool which they see as the future of mental health care.”
Michaels explained how the landscape of mental health care has shifted in a 2025 article published by The American Psychoanalyst . “During the pandemic, the imbalance between supply and demand in mental health care — low supply of therapists at a time of high demand for therapy — caught the attention of financiers,” Michaels wrote. “They had previously been making profitable inroads into medical/surgical health care, buying up outpatient surgery sites, MRI centers and physician practices. They then turned their attention to mental health, buoyed by the supply and demand imbalance and the presence of government and insurance monies that paid for therapy.” In short order, mental health practice management companies began popping up that prioritized slashing expenditures, a job made possible by cutting payments to therapists, upping their workload and regularly auditing them for alleged compliance violations. At the same time, a slew of online therapy platforms emerged and began marketing themselves as 24/7 tools to supplement — or completely replace — therapy with trained practitioners. “We know nothing about the long-term efficacy of AI therapy.” It’s not just patients who have been targeted by these startups. Extensive marketing campaigns have successfully lured therapists around the country to affiliate with app-based programs. Kate P., a licensed marriage and family therapist in Southern California, told Truthdig that during the 2020 COVID shutdown, she moved her practice completely online, a transition she describes as relatively straightforward. “Things went along fine until about 2023, when all of a sudden I started getting aggressive marketing messages from mental health companies like BetterHelp , Alma , and Headway ,” she says. Among other things, they promised to handle billing, credentialing with insurance companies, scheduling and note-taking during sessions. “It sounded so convenient that I signed up with Alma and pay $125 a month as a membership fee.”
Several months in, however, Kate realized that she had been sold a shiny, but false, bill of goods. Moreover, she realized that she had not only failed to ask about Alma’s cut from each of her sessions, but she had not locked in the per-session fees she’d been promised. She was subsequently gobsmacked when her fees were reduced.
Kate is now trying to get off the Alma platform, but will have to be re-credentialed by each insurance company if she wants to accept their coverage. Her only other option is to solely treat people who can pay for therapy out of pocket — an alternative that an increasing number of providers are choosing. Like Kate, Shawn FitzWilliam , a Missouri counselor, quickly became disenchanted with Alma after he learned that the platform was taking two-thirds of the fees paid by his clients for treatment. He also bristled when he received a notice informing him that he would only be paid for sessions lasting between 32 and 45 minutes. “Alma blames [the insurer] Aetna for the change, but this amounts to a 20% pay cut for therapists and Alma did nothing to fight the reduction,” FitzWilliam says. “They’re also requiring therapists to upload all of our notes for internal review without telling us how this information will be used and are requiring us to use their portal to schedule sessions with clients.” FitzWilliam also questions Alma’s claim that it is HIPAA compliant. He says these reasons, and a general lack of operational transparency, led him to end his contract with the company. “I’m still using Headway,” he says, but potential conflicts there could undermine his ability to work with clients as he sees fit. “I can opt out of their AI note-taking, but if I do, I’ll lose ‘clawback’ protection if the insurers later decide that I’ve been overpaid,” he says.
This stranglehold is compounded by the constant marketing of these platforms to therapists and potential users. The number of platforms that offer AI mental health support is mushrooming. Among them are EBB (“EBB can help you unpack what’s on your mind and move forward with more clarity”), HeadSpace Care (“Mental health support. Anytime. Anywhere. 24/7 for everyday support”) and WYSA . One provider, 7 Cups (“Caring listeners and affordable online therapy and coaching”), bills itself as the world’s largest mental health community, with AI companion “Noni” and 563,378 “trained listeners” reaching more than 72 million people throughout the world. The number of platforms that offer AI mental health support is mushrooming. Clare V. is a Massachusetts therapist who specializes in treating trauma survivors, teens and young adults. “Chatbots are obviously cost effective,” she says. “They don’t need breaks, meals or healthcare and never take a vacation or sick day, but they’re also not human.” She decided to check out ChatGPT for mental health, posing a 25-year-old with suicidal depression. The bot’s response appalled her. “I was told to take a shower and spend time with friends or family,” Clare says. “That’s all it kept saying. It made me wonder if anyone can be held accountable if someone jumps off a bridge.”
Clare describes the experience as troubling. “I work with a lot of kids who are in gender transition and I have a safety contract with each of them. I make clear that if they are feeling suicidal they can call or text me 24/7. I want them to reach out to me rather than to a robot that will tell them to take a shower. Chatbots can’t hear nuance in a voice. They can’t see body language. They’re literal. I don’t even use AI to record sessions because I can take notes all by myself.” Christopher AhnAllen , director of psychology enrichment at Mass General Brigham hospital, says that thanks to the shortage of mental health workers, he understands why people who want help turn to AI. “Some people also use it to augment what a licensed human therapist can provide,” he says. “But the potential for AI to offer information that conflicts with what the human therapist has said is a serious risk.”
“A client of mine used an AI chatbot to discuss a work situation but was given information about the workplace that was untrue,” he says. “This led to psychiatric distress, suicidal thinking and a hospitalization. When we finally learned what the robot told the person, we were dumbfounded and still can’t figure out where the misinformation came from. This person was duped by AI.”
Nonetheless, AhnAllen sees some potential good coming from online platforms. “Going online may reduce the stigma of seeking mental health services, especially for kids who may not want to ask to see a school counselor or tell their parents they want to see a therapist,” he says.
Insurance companies embrace AI In addition, as AI and online therapy become normalized, the insurance industry is making it harder for therapists to get paid for the care they deliver — while simultaneously promoting AI services to clients. Brianna Reinhold , a licensed professional counselor, is the founder of Northern Lights Therapy, a three-center psychotherapy practice in Arizona. She is currently in a battle with Blue Cross Blue Shield over rules governing coverage and reimbursement for treatment.
“We’re a one-stop shop for people of all ages, and we serve individuals, families, couples and first responders,” she told Truthdig. In mid-December 2025, Reinhold learned that Blue Cross was changing its billing procedures, and practitioners would have to follow new rules as of Jan. 1. Under the new plan, Reinhold would have to personally handle the initial intake process for each new client, create a treatment plan based on the interview and offer a diagnosis. She would also have to be physically present in the clinic when clients were seen by the 15 therapists she employs. The insurance industry is making it harder for therapists to get paid. “Previously, I would review everything, but I did not need to be on-site during every session. The new rule meant I would have to close the centers if I took a vacation or was sick,” she explained. “As clinicians, we pushed back. Blue Cross finally agreed that I don’t have to be present during every session but they continue to demand that I do all the intakes and develop all the treatment plans and diagnoses.” In early January, Reinhold reached out to the media about the changes; a little more than a month later, she received an audit notice. “They requested files on every Blue Cross client we’d seen between October 2022 and October 2025. This is hundreds of people because Blue Cross covers 70% of our caseload.” In May, Reinhold got the audit results: a “clawback” asking for $298,000 in alleged overpayments for care that the insurer said was not “medically necessary.” Since then, Blue Cross has asked to review the medical records of more than 50 additional clients. Reinhold’s clinics are not the only facilities impacted by the rules change: Three other Arizona mental health practices in addition to Reinhold’s have been audited, with similar results. All told, Reinhold says, the centers are on the hook for a combined total of $1.4 million.
“We’re now doing a $100 pay rate per session and seeing some people pro bono,” Reinhold says. “Meanwhile, some of our clients have received letters referring them to a Blue Cross virtual platform for AI mental health services. They’ve been offered three sessions with no copay.”
“The pace of this is mind-blowing,” she says. “I’ve gone back on the news because I want people to be forewarned, but our lawyers tell us that this will likely be a long fight.”
Reinhold is now part of the Mental Health Insurance Reform Task Force, a national coalition of providers who are challenging the insurance industry’s tightening restrictions on providers of care and reduction of payouts for claims.
David Lloyd, chief policy officer at Inseparable , a national mental health advocacy group, works at the state and federal levels to improve access to treatment services. “AI is developing so rapidly that states are struggling to respond,” he says. “But since we know that AI is being used for mental health support, it is important for us to put guardrails in place, particularly to protect youth from the worst aspects of it.” “AI is developing so rapidly that states are struggling to respond.” Lloyd notes that several states have already taken small steps to rein in AI : Colorado recently passed House Bill 25-1002 , which bans clawbacks by insurers for past payments and restricts AI in clinical settings to administrative functions that are then reviewed by a licensed human provider. Illinois and Nevada also prohibit therapists from using AI for diagnostic purposes and, like Colorado, limit its use to administrative functions. New York and Utah now require AI companies to disclose that “companion chatbots” are machines, and Delaware has banned health insurers from demanding prior authorization for mental health care. Bills to restrict AI in mental health settings or regulate clawbacks are currently pending in California, New Jersey and Pennsylvania.
“The insurance and tech companies are on a different page than therapists and other care providers,” Michaels of the PSiaN says. “When people sign up to join a mental health practice management company like Alma, they typically have no idea who owns the company or who has invested in it. They’re shocked to learn that these companies are often owned by insurance companies.”
To wit, UnitedHealthcare Group, through its subsidiary Optum, owns both the insurance and provider sides of care, and private equity firms like the Blackstone Group and KKR are starting AI psychotherapy and management groups.
It’s clearly profitable. The global market for mental health apps was valued at $7.4 billion in 2024 .
Kate P., the Southern California therapist who is trying to leave Alma, finds this outrageous, and calls it a betrayal of her field’s mission. “Being a therapist is an incredible job,” she says, one that AI will never be able to do. “I get to be a witness to what it is to be human and actually help people in ways that matter.”
The post AI’s Threat to Mental Health Care appeared first on Truthdig .