Samwise Healthcare IT Newsletter
Friday, October 2, 2026
California Governor Signs Three Bills Governing Healthcare AI, Vetoes Two
California Gov. Gavin Newsom signed three bills this week governing healthcare AI, Healthcare IT News reported Oct. 1. AB 1979, sponsored by the California Nurses Association, limits AI to an advisory role and prevents it from performing licensed clinical functions. SB 503 requires developers and deployers to make reasonable efforts to mitigate known or reasonably foreseeable risks of biased impacts from clinical decision support systems. AB 1609 bars large enterprises from presenting customer service chatbots as human, exempting hospital communications. Newsom vetoed AB-2575, which would have added employment protections for healthcare workers using AI, and SB-903, which would have barred AI from mental healthcare therapeutic functions.
Sources: Healthcare IT News
29 Health Systems Affected by Oracle Health Data Breach
Becker’s Hospital Review updated its list of hospitals and health systems reporting that patient data was compromised in a 2025 breach of Oracle Health’s legacy Cerner systems; the list, updated Oct. 1, now names 29 organizations. The hack occurred as early as Jan. 22, 2025, and health systems continue to alert patients. Oracle Health asked healthcare organizations to delay notification during the investigation, and the EHR vendor faces legal action over the incident. Affected organizations on the list include Atrium Health, AdventHealth, Christus Health, OSF HealthCare, ChristianaCare, LifeBridge Health, Baptist Health South Florida, Methodist Le Bonheur Healthcare, Munson Healthcare and Union Health in Terre Haute, Indiana.
Sources: Becker’s Hospital Review
ARPA-H Awards UVM Up to $38M for ICU Digital Twin AI
The Advanced Research Projects Agency for Health has awarded the University of Vermont’s Larner College of Medicine up to $38 million over five years to develop AI-powered digital twins for intensive care units, Healthcare IT News reported Oct. 1. The milestone-based ReSCUED project, funded under ARPA-H’s CIRCLE program, targets severe trauma, burns and sepsis. Researchers will collect patient blood every six hours, model immune responses and use AI to suggest treatments tested virtually on each patient’s digital twin. Partners include Wake Forest University, the University of Alabama at Birmingham and Washington University. UVM says digital twins could shorten ICU stays by as much as 25%.
Sources: Healthcare IT News
UC San Diego Health, Anthropic and Harvard Build Oversight System for Patient-Facing AI
UC San Diego Health is working with Harvard Medical School and Anthropic to develop an open-source oversight system to check the safety of medical recommendations made by patient-facing AI, Becker’s reported Oct. 1. The collaboration is part of ARPA-H’s ADVOCATE program, which focuses on AI-supported care for patients with heart failure and aims to develop an FDA-authorized agentic AI system for cardiovascular care. The approach centers on a “clinical constitution,” a set of rules to guide safety assessments. “In the long run, simply evaluating an AI tool before it goes live is not going to be enough,” said Dr. Karandeep Singh, UC San Diego Health’s chief health AI officer.
Sources: Becker’s Hospital Review
Nearly 12 Million People Live in Broadband Deserts That Limit Telehealth
Nearly 12 million people across 41 states live in areas with inadequate broadband, limiting telehealth’s potential, according to a University of Vermont study in JAMA Health Forum covered by Healthcare IT News on Oct. 1. Of those residents, 10.5 million, or 88%, live in rural areas, and the South had the largest affected population at about 6.5 million. Nearly 650,000 rural residents face simultaneous gaps in broadband, nearby ambulance services and healthcare facilities. Researchers defined broadband deserts as areas below 100 Mbps download and 20 Mbps upload speeds. The authors called for targeted infrastructure investments and expanded reimbursement policies to support virtual care.
Sources: Healthcare IT News
Privacy Concerns Hold Patients Back From Using AI, Survey Finds
Patients are less willing to use artificial intelligence for healthcare when it requires access to personal data or protected health information, according to a Wolters Kluwer survey reported by Healthcare Dive on Oct. 1. Ipsos polled 254 U.S. patients in March for the survey. Around 89% said AI responses in healthcare need human oversight, and over 70% worry about the privacy of their protected health information. Patients also cited concerns about bias (72%) and hallucinations (69%). Only about 30% said they would use AI to find a healthcare provider, and 19% to navigate insurance coverage. Nearly half said AI health advice tools should undergo federal testing and approval.
Sources: Healthcare Dive
Ortet Launches Health AI Lab with $500M Commitment From Thoreau
AI startup Ortet has launched with a $500 million commitment from health technology investment platform Thoreau to build AI models designed to connect clinical, biological, operational and financial health data, MobiHealthNews reported Oct. 1. The New York-based company is led by cofounder and CEO Kyunghyun Cho, an AI researcher who previously established the frontier research team at Genentech. Ortet plans a patient-centered foundation model adaptable across scientific research, clinical care, healthcare operations and financial administration, and says it has already secured its initial GPU cluster. The commitment will support compute and data infrastructure, model training and deployment, and research talent. Ortet is one of Thoreau’s founding investments.
Sources: MobiHealthNews
Payerset Debuts AI Assistant Built on Price Transparency Data
Payerset launched an AI-powered research assistant that unifies negotiated rate data, claims, policy and market benchmarking so health systems can ask plain-English questions about their market position, Fierce Healthcare reported Oct. 1. The platform draws on 20 trillion rate records retained since 2024, with more than 10 trillion more processed every quarter, according to the company. The tool is available as a Model Context Protocol server that plugs into enterprise AI tools including Claude and ChatGPT, and as a chat-style portal. Payerset has nine hospital and health system customers, including Northwell Health, WakeMed and Cone Health, and says it has been profitable since it started.
Sources: Fierce Healthcare
Every newsletter preserved and searchable
Curated by JD · samwise.agency