A new American Psychological Association survey of 1,242 psychologists finds 77% have patients discussing AI chatbot use in therapy, with 97% worried these bots can reinforce delusions or negative thinking instead of challenging it.
Character Technologies and Google have agreed to settle wrongful-death lawsuits from families in four states who say Character.AI chatbots contributed to their children's suicides, including the high-profile case of 14-year-old Sewell Setzer III.
The bipartisan GUARD Act, which would ban minors from AI companion chatbots and require age verification plus steep penalties for harmful bots, cleared the Senate Judiciary Committee unanimously and now awaits a full Senate vote.
At its second annual London Innovation Summit on October 7, 2026, Cleveland Clinic disclosed a new AI research pilot with IBM to study how routine surgeries affect patients' long-term health and quality of life, drawing on imaging and clinical data from its London BioResource. The announcement came alongside summit sessions on AI-enabled robotics and the 'connected operating room.'
Israeli startup Sleep AI, built on Technion sleep-research, says a home finger-oximeter paired with an AI model can diagnose sleep apnea with up to 99% accuracy for moderate-to-severe cases, at a fraction of the cost of an overnight sleep lab. The device has not yet won regulatory approval, and its figures remain company-reported rather than independently validated.
New York's Office for the Aging says its ElliQ companion-robot pilot cut loneliness among 800-plus seniors by 95%, with users engaging more than 30 times a day. The figure comes entirely from state and vendor data rather than an independent study, even as the program heads into a second year.
NewYork-Presbyterian has adopted Signal 1's AI Management System to govern every AI tool across the hospital network under one oversight platform, reflecting a shift from vetting individual AI products to managing a growing portfolio of predictive, generative, and agentic AI tools.
NHS England South East is rolling out Skin Analytics' DERM AI tool across all 18 regional hospital trusts, its largest autonomous-AI deployment to date, reaching about nine million more people. The system can discharge patients with benign lesions without a specialist visit, but regulators have kept its NHS approval conditional while more evidence is gathered.
A Mount Sinai-led study published in Nature Medicine found ChatGPT Health under-triaged 52% of true medical emergencies, including cases of diabetic ketoacidosis and impending respiratory failure. OpenAI disputes the study's real-world relevance, while outside physicians say the tool isn't yet reliable enough for urgent-care decisions.
GE HealthCare and Mass General Brigham announced a research collaboration on September 23, 2026 to explore generative AI that helps radiation oncology teams navigate scattered treatment records. The project builds on earlier MGH software that GE says cut treatment start times from 30 days to eight at one hospital.