Samwise Healthcare IT Newsletter
Friday, July 25, 2026
White House orders HHS, NIH to prioritize AI in FY2028 budgets
A White House memorandum issued July 21 directs HHS, NIH, and every other federal department to realign fiscal year 2028 research and development budgets around nine priority areas, led by artificial intelligence and foundational biological sciences. Accompanying the order is an Office of Science and Technology Policy report titled “Science: A New Golden Age,” which Director Michael Kratsios described as the first comprehensive rewrite of U.S. science policy in more than 80 years. The Genesis Mission — President Trump’s AI-for-science initiative — is designated a central budget priority. All federal agencies with at least $3 billion in FY2026 R&D funding must submit implementation plans within 90 days.
Sources: Becker’s Hospital Review Share ↗ ✉︎ Email 💬 Text
Healthcare data breaches climb to 281 in the first half of 2026
Healthcare data breaches climbed to 281 incidents in the first half of 2026, up from 270 in the same period of 2025, according to a new report. Hacking and IT incidents are the leading cause, accounting for most exposures. The year-over-year increase continues a sustained trend of rising cyber incidents against the healthcare sector. The numbers signal that health systems and their business associates continue to face persistent pressure from threat actors targeting patient data and clinical systems. For security and IT leaders, the report reinforces the urgency of ongoing investment in defenses as the pace of incidents in the first half outpaces the same period a year ago.
Sources: Becker’s Hospital Review Share ↗ ✉︎ Email 💬 Text
Ardent Health tops 1 million patient encounters with ambient AI
Ardent Health has surpassed one million patient encounters using ambient AI, deployed through a partnership with Ambience Healthcare. The technology covers 87 percent of the health system’s ambulatory footprint. Clinicians using the tool report saving more than three hours per week on documentation. One physician — Dr. Zheng — reduced documentation time by 53 percent. The milestone reflects rapid maturation in ambient AI adoption as health systems move from pilot phases into enterprise-wide deployment. Ardent’s results add to a growing body of evidence that ambient documentation delivers measurable time savings for physicians and eases the administrative burden that has long been a driver of clinician burnout.
Sources: Becker’s Hospital Review Share ↗ ✉︎ Email 💬 Text
OpenAI expands ChatGPT Health to all U.S. users aged 18 and older
OpenAI has expanded ChatGPT Health to all eligible users in the United States aged 18 and older, following a more limited initial rollout. The feature lets users securely connect personal health records — including Apple Health data and records from connected hospital systems — so ChatGPT can provide personalized health information. The tool runs on OpenAI’s GPT-5.5 and GPT-5.6 models. The company states that health data is not used to train its models. The expansion marks a significant step toward making AI-powered personal health tools broadly accessible and signals OpenAI’s growing ambitions in the consumer health market.
Sources: Becker’s Hospital Review Share ↗ ✉︎ Email 💬 Text
Johns Hopkins builds benchmarking harness to evaluate AI agents before deployment
Johns Hopkins Medicine has developed actAVA X-BENCH, a benchmarking framework for evaluating AI agents before clinical deployment. Testing found that agents complete only 28 percent of complex healthcare tasks on their first attempt, with fewer than 8 percent performing consistently across repeated trials. The framework evaluates agents against NIST and HIPAA governance standards, surfacing failure modes that are rarely visible in standard vendor demonstrations. Researchers argue that harness design — the structure of the testing environment itself — is as consequential as model selection in determining real-world reliability. The work signals a broader push toward rigorous standardized evaluation before health systems move from AI pilots to agentic workflows.
Sources: Healthcare IT News Share ↗ ✉︎ Email 💬 Text
'Who should be monitoring this?' Health systems grapple with post-deployment AI oversight
AI governance at health systems often breaks down after go-live, with responsibility for monitoring model performance quietly becoming ambiguous. A Becker’s report draws on leaders at seven major health systems who say the question — who should be monitoring this? — has no clear answer. Henry Ford Health CDIO Geoff Patterson called treating go-live as “the finish line” the most common governance failure. Hackensack Meridian Health now requires documented outcome metrics before any AI pilot can expand. Sentara CIO Tim Skeen described performance instrumentation as a “build-time requirement.” Penn Medicine, Intermountain Health, and Sanford Health say post-deployment monitoring governance remains a work in progress.
Sources: Becker’s Hospital Review Share ↗ ✉︎ Email 💬 Text
Teladoc Health launches Teladoc One whole-person virtual care model
Teladoc Health has launched Teladoc One, a new virtual care model designed to replace fragmented, condition-by-condition care with a whole-person approach coordinated across conditions and settings. At its core is Pulse, a clinical engine that coordinates care across care pathways. The model shifts the compensation structure, tying fees to patient outcomes rather than individual visit volume. A pilot with select health plan partners is scheduled for September 2026, with a broader rollout planned for January 2027. The launch signals a strategic pivot for Teladoc as it seeks to differentiate in a competitive virtual care market by emphasizing longitudinal, value-based care.
Sources: Becker’s Hospital Review Share ↗ ✉︎ Email 💬 Text
Rhode Island and Hawaii enact healthcare AI laws; Arizona governor vetoes similar bill
Rhode Island and Hawaii have each enacted laws governing AI chatbots and mental health AI applications, establishing guardrails around their use in clinical and patient-facing contexts. Arizona considered comparable legislation but the governor vetoed the bill, illustrating divergence in how states are approaching healthcare AI governance. The patchwork creates a growing compliance challenge for health systems operating across multiple states, as there is no uniform federal framework to guide implementation. Regulatory and legal teams at national organizations are now tracking state-by-state developments as legislatures increasingly step into a space where federal rulemaking has moved slowly.
Sources: Becker’s Hospital Review Share ↗ ✉︎ Email 💬 Text
How Sentara Health built an enterprise AI literacy program with 60,000+ training modules
Sentara Health has built an enterprise AI literacy program that has delivered more than 60,000 training modules across its workforce. The program includes an AI Trailblazer Certification designed to credential employees who demonstrate proficiency in applying AI tools within clinical and operational workflows. Teams completing training have documented specific workflow improvements, giving the health system evidence that AI adoption is translating into measurable operational gains. The initiative reflects a deliberate strategy to build foundational AI knowledge before scaling deployment, addressing a common gap where technology adoption outpaces the readiness of the workforce expected to use it.
Sources: Healthcare IT News Share ↗ ✉︎ Email 💬 Text
Curated by JD · samwise.agency
Leave a Reply
You must be logged in to post a comment.