Technology news for families navigating serious mental illness — translated so you don’t have to.
Four stories this week, all circling the same question in different ways: who’s actually watching the tools we’re leaning on?
A warning sign worth knowing about
A recently recirculated case report is worth knowing even though it’s not new: a 60-year-old man with no psychiatric history asked ChatGPT for a healthier alternative to table salt. It suggested sodium bromide — a compound that resembles salt but builds up in the body over time and can cause hallucinations, paranoia, and psychosis. He used it for three months before landing in the hospital on an involuntary psychiatric hold. Live Science has the details. Worth being precise about what this is and isn’t: it’s not the same thing as a chatbot reinforcing an existing delusion — this was straightforwardly bad medical advice with a toxic, physical cause. But the lesson for any family is the same either way: a confident-sounding answer isn’t the same as a correct one, especially for anything you’d normally ask a doctor or pharmacist first.
The most useful signal might be silence
Psychiatric Times reports on a genuinely counterintuitive finding from Amae Health’s chief medical officer: one of the strongest early-warning signs for relapse isn’t anything a smartwatch measures — it’s when someone simply stops wearing or charging it. Rather than treating that gap as a technical failure, researchers are now building it directly into predictive models as a meaningful signal on its own. If a loved one who’s usually consistent about wearing a device suddenly stops, that’s worth a check-in, not just a shrug.
California pushes back — with real teeth
A new bill in California, SB 903, would bar companies from advertising chatbots as therapy and require a licensed professional to review any AI-driven therapeutic decision. What makes this one different from similar efforts elsewhere: it’s arriving alongside a formal complaint, filed by the National Union of Healthcare Workers, against Kaiser Permanente over its alleged use of an algorithm to triage mental health patients. Regardless of how that complaint plays out, it’s a sign this isn’t an abstract policy debate anymore — it’s happening inside one of the country’s largest health systems right now.
When “better than nothing” becomes the whole plan
Related, and worth reading alongside the Kaiser story: a Forbes piece makes the case that insurer- and employer-funded AI mental health tools may be quietly squeezing the human therapists they’re supposed to support — as reimbursement stays flat and more clinicians drop out of insurance networks altogether, “AI is better than no care” risks becoming the permanent answer for anyone who can’t afford to go around it. I’m actually working on a longer piece exploring this exact tension — more on that soon, and if it’s accepted, I’ll share it here first.
Come visit us and keep abreast of the latest & greatest technology news that impacts the world of serious mental illness.
