Florida sued OpenAI, alleging ChatGPT was deployed without proper warnings and linking the chatbot to two deadly shootings — a first-of-its-kind legal test of platform liability that could reshape how mental‑health risk is managed around AI (Fox5). [P]At the same time, a RAND survey finds nearly 1 in 5 young people now turn to AI chatbots for emotional support, amplifying concerns about unregulated therapeutic use (NBC News). On the treatment front, a phase‑2 trial of the psychedelic luvesilocin showed rapid durable benefit for postpartum depression, offering a potential fast-acting option amid widening gaps in care (Medscape).

California State University's $39 million no‑bid deal to give 470,000+ students ChatGPT Edu has students and faculty up in arms, spotlighting corporate influence on campuses (WebProNews). [P]The American Federation of Teachers' Randi Weingarten pushed a 10‑point plan calling for elementary screen bans and AI limits, crystallizing a new battleground over how tech is taught and restricted in early grades (WebProNews). Universities and nonprofits are trying to catch up — Boston University and Just Horizons Alliance are building an AI Ethics Index for schools to set guardrails while policy and pedagogy scramble to keep pace (Axios).

Connecticut’s attorney general opened a formal probe into Roblox over child safety practices, a reminder that virtual playgrounds are now frontline parenting concerns (NH Register). [P]New Mount Sinai research links a strong genetic component to postpartum psychosis, which could transform screening and supports for new parents at high risk (EurekAlert). And a 17‑year‑old learner’s permit driver who crashed into Venice Beach Pavilion at 70 mph underscores the continuing, very real-world duties of supervision and oversight in teen driving (FOX13).