Unreasonable Adjustments
Just when you think our mental health can’t get any worse the next circle of behavioural healthcare hell opens up. Over the last week Donald Trump has signed the form that expands an AI pilot by Limbic to provide AI therapy across federal health care systems in the States. If you’re on Medicare, you will rapidly only be able to access automated care. For every 1000 consumers treated by an ‘AI Analyst’ they will be overseen by one actual clinician. No longer will digital entrepreneurs have to wade through the red tape of developing medical devices and jump through clinical hoops, just a series of general AI agential decisions that nobody can effectively either understand or regulate.
In the gaslit language of digital ecommerce, this Limbic AI pilot scheme offering effective and scalable strategic solutions to chronic condition management, is called…..wait for it…… ACCESS.
Just when you think you know every dark corner of Angerland the penny drops that this model allows the widespread automation of therapy and the introduction of ‘outcome aligned’ payments where providers are paid more if their patients feel better. Even if you could get the data about risks and harms of this model of AI therapy published in the current academic publishing system, there wouldn’t be anything you could do about it.
This year Alma, one of the largest clinical practice management platforms in the US, unilaterally announced that therapists in private practice with clients who have private medical insurance with Aetna will see what has become known as ‘down-coding’ of therapists’ capacities, their session times reduced and a subsequent 40% cut in pay. Although Alma is not owned by Aetna, it’s business model relies on being a contractor to the large private medical insurance companies who determine the parameters of therapeutic practice. In the same week Aetna announced its new service, Mental Health On Demand, offering single session therapy with non-clinical staff or, marketed as the emerging preference of consumers, an AI agent who will respond within 13 seconds. The cuts at Alma represent a loss of income for therapists ranging from a drop of $4800 a year for 5 Aetna sessions a week to $14400 for 15 sessions. Private providers now determine what counts as therapy.
A series of important collective legal and regulatory mobilizations have emerged through the National Union of Healthcare Workers (NUHW) and Psychotherapy Action Network (PsiAN) - organizing organizations that are doing the work.
They are the first to understand down-coding and the subsequent gold rush for automation of therapy. Instead of wondering whether Claude was useful they made formal complaints to the San Francisco Supervisors about what AI roll out means across the private medical insurance sector, by the big players at Kaiser Permanente. If you want to understand the playbook of how automation takes place across therapy this is mapped out in all its granular glory in the testimonies at this hearing from NUHW members and Linda Michaels from PsiAN.
NUHW’s current legal campaign to secure regulation in California - Regulation SB 903 - is an important attempt to protect both clinicians and their clients operating within the UberTherapy business model. SB 903 demands explicit consumer consent for recording therapy sessions and no penalties for clients who say no. It demands that clinicians not AI Agents carry out triage or make clinical decisions. It puts clinicians firmly in the role of monitoring AI outputs, and makes it illegal to blur the line between a chatbot and an actual human therapist. It also sets to stop AI mental health providers from sharing or selling private data to make even more money.
The other people who might understand deeply the implications of UberTherapy are Uber Drivers. Worker Info Exchange is now coordinating with app driver unions around the introduction of AI technologies that are behind the era of dynamic pricing and pay on platforms. The technology, outlined in Worker Info Exchange’s current collective action against Uber as part of their dynamicpay.org campaign, explains the material and non-material harms that are at the heart of the next generation of robofiring technologies. Except this time it’s not even an algorithmic management system rolled out by Human Resources, it’s an AI agent that nobody actually understands making the decisions. You can listen to our conversation with James Farrar from WIE about how to prepare for dynamic pay in therapy.
It’s just unreasonable when you think about it.
Our next project is called MuchBetter.Help to encourage our own and each others’ unreasonable adjustment to this unreasonable reality. To imagine a UK dynamic pay campaign and a UKSB 903 legal mobilization for therapy. The first submission to the MuchBetter.Help Library is a series of podcasts Stories of UberTherapy, available free to listen on MuchBetter.Help and a podcast platform of your choice
During 2026/2027 with colleagues from the US and UK, we will be piloting, designing and publishing useful information and workplace tools for navigating this new terrain.
Follow the story of MuchBetter.Help here
If you’re an UberTherapist join us here