Check out the new SODP website and be sure to visit the always-expanding SODP + JMIR library of expert webinars. All webinars are free to view and approximately 25 minutes long. Are you interested in being on our next webinar? If so, please reach out with your proposed topic.
By all metrics, 2026 is on track to be one of the busiest years for digital mental health as interest in AI fuels both patient and clinician interest. The Society of Digital Psychiatry and JMIR Mental Health will soon roll out Digital Navigator training to help you or your team best utilize AI in care settings. We will also soon have a first release of the mindBench.ai evaluation framework for AI, offering you hands-on support for understanding and selecting the AI that best matches your interests and needs. Finally, to complete our three-pronged roadmap, we will, of course, continue our monthly webinars but are planning an expanded annual symposium for Fall 2026.
📈Trends in Digital Mental Health What is Real and Does That Matter?
As more and more research on AI for mental health is produced, it feels impossible to keep up. Perhaps the solution is to use AI to read the papers and summarize them, completing a concerning cycle if one is also worried that many of these same papers may have been written in part by AI. To break free of this loop, it is worth considering two broad types of AI and mental health studies: those about simulated people and those with real people.
Human behavior is hard to predict, and mental illnesses are even more challenging to chart. Clinical guidelines, diagnoses, and treatment pathways offer us tools to help steer care, but anyone who has worked clinically will agree that there is rarely a single patient who fits perfectly into any guideline, diagnosis, or treatment. One of the factors that makes clinical work so rewarding is that we constantly learn from our patients, who teach us new facts about their condition and how to treat it. Returning to AI, we often find the research replete with impressive studies with large samples but in reading closely, we see that the large sample is either a computer simulation of what said computer thinks people would do/act, people pretending to be patients and talking/typing as they think said patient would, or people acting out scripts and roles written by computers. We can, of course, learn much from this research, and it is best to consider it pre-clinical developmental work, using an analogy from pharmaceutical drug development. This is where we learn about if the AI model has a chance to work, can be safe, and perhaps even engaging. It is critical and foundational work. But what happens when real people, be those patients or clinicians, use the AI in the real world?
There is less research on such real-world use cases, but this is arguably the most relevant research for the clinical mental health community. We learned from a decade of mental health apps that simulating use (i.e., asking people if they would engage with mental health apps) yielded the wrong answers…. real-world engagement is rather low. A now-famous (and infamous to some) 2019 study of publicly available app data found that the median retention for mental health apps after 15 days was under 4% [1]. AI is not the same as apps, and simulation of AI is not the same as self-assessment of engagement. Still, a 2026 study examining teens' objective use of AI (though monitoring their phone use) found that average use was ~3 minutes per day.
In part, there are so many AI studies because so many of them rely on simulated data. In the time it took you to read this sentence, I could have simulated 1 million patients. Thinking through which clinical questions demand real clinical samples (be that patients or clinicians) and which ones may not can help you sort through the hundreds of new AI and mental health papers that likely were published today!
Author’s note: While AI was not used to write this piece, I did ask AI if it could simulate that many patients. It assured me the real number is actually 6 million and then to make me feel better told me that it is a ‘very good question.’
1. Baumel A, Muench F, Edan S, Kane JM. Objective user engagement with mental health apps: systematic search and panel-based usage analysis. Journal of Medical Internet Research. 2019 Sep 25;21(9):e14567. 2. Maheux AJ, Akre-Bhide S, Boeldt D, Flannery JE, Richardson Z, Burnell K, Telzer EH, Kollins SH. Generative Artificial Intelligence Applications Use Among US Youth. JAMA Network Open. 2026 Feb 2;9(2):e2556631.
📖 Book Spotlight
A Practitioner's Guide to Advancing Behavioral Health Care With Artificial Intelligence by David D. Luxton PhD delivers on exactly what the title promises. The book is thoughtfully written to be relevant well beyond the current generation of AI and provides practical tips and best practices for working with AI. Learn more.
👍 Survey Spotlight
A team at the University of Basel wants to learn about how clinicians use AI as part of psychotherapy. They are asking for the support of any SODP members who are trained psychotherapists, psychologists, counsellors, psychiatrists, and other mental health professionals.
Key survey details:
🕛 5–7 minutes 🌏 Global survey 🗣️ English or German 🔒 Completely anonymous 📆 Closes soon
Join the JMIR and SODP team in person at the American Psychiatric Association Annual Meeting in San Francisco on May 16-20. Stop by booth 1051 for free giveaways and fun surprises (of an educational nature).
While not often thought of as a digital mental health meeting, the American Society of Clinical Psychopharmacology annual meeting in Miami, May 26-29, really is. The meeting title says it all: ‘Psychiatry and AI: Promises, Perils, and Pathways Forward’.
🌐 Upcoming Webinar Spotlight
March 24, 2026, 3 PM ET
Join JMIR Mental Health and the Society of Digital Psychiatry for a live virtual webinar Bridging Innovation and Care: Evaluating Digital Health Navigators in Prescription Digital Therapeutics on March 24, 2026 at 3 pm ET, featuring Julian Schwarz, MD. Dr. Schwarz is a SODP co-lead and a clinician-researcher at Brandenburg Medical School in Rüdersdorf, Germany.
Get your brand in front of highly engaged clinicians by advertising with SODP. In addition to our membership journal newsletter, we offer a society website and access to more than 32,000 online professionals. Email Kathryn Ryan to learn more about running your first campaign with us.
-1.png?upscale=true&width=1200&upscale=true&name=Untitled%20(1000%20x%20400%20px)-1.png)
