Samwise Healthcare IT Newsletter
Thursday, July 17, 2026
Healthcare ransomware attacks surge 14% amid growing cyberthreats
Healthcare ransomware attacks reached 410 incidents in the first half of 2026 — a 14% jump over the same period last year — according to a new FortiGuard Labs report. That translates to roughly 2.3 attacks every day, with 247 provider organizations targeted. The most active groups include Qilin, The Gentlemen, LockBit, and INC. Compounding the sector's risk, the HIPAA Security Rule final rule — which would mandate new cybersecurity baseline controls — has been delayed again, with enforcement now pushed to July 2027. Security experts warn that the delay leaves patient data vulnerable as attackers grow more sophisticated.
Sources: TechTarget / HealthTechSecurity Share ↗ ✉︎ Email 💬 Text
Trump admin introduces Gold Eagle AI initiative to find and fix cyber vulnerabilities
The Trump administration launched "Gold Eagle" this week, an AI-powered cybersecurity clearinghouse designed to identify and remediate vulnerabilities across critical infrastructure, including healthcare. National Cyber Director Sean Cairncross and DHS Secretary Markwayne Mullin announced the initiative, framing it as a proactive effort to stay ahead of state-sponsored and criminal threat actors. Gold Eagle will aggregate threat intelligence from federal agencies and private sector partners, using AI to correlate signals and prioritize remediation. Healthcare organizations — among the most targeted sectors — are expected to be early beneficiaries of the shared intelligence platform.
Sources: Healthcare IT News Share ↗ ✉︎ Email 💬 Text
AI in health care: The governance gap
Healthcare organizations racing to deploy AI face a widening governance gap, according to attorney Christine Chasse of Spencer Fane. The DOJ recovered $6.8 billion under the False Claims Act in fiscal year 2025 — $5.7 billion from healthcare — signaling regulators are watching AI-assisted billing closely. Kaiser Permanente affiliates reached a $556 million settlement tied to unsupported Medicare Advantage diagnosis coding. The FDA issued revised clinical decision support software guidance in January 2026; the EU AI Act imposes penalties up to €35 million or 7% of global annual turnover. Pennsylvania has also sued Character.AI over chatbots alleged to have posed as physicians.
Sources: Health IT Answers Share ↗ ✉︎ Email 💬 Text
CHAI launches PULSE AI adoption program for public health agencies
The Coalition for Health AI (CHAI) announced the PULSE program, an initiative to accelerate AI adoption among the nation's roughly 3,000 local public health departments. The program will support approximately 2,000 practitioners across five initial AI use cases — a significant push given that only 5% of local health departments reported using AI in 2024. OpenAI and Anthropic have donated software licenses to support the effort. Pilots are slated to begin in fall 2026, with PULSE focused on helping departments integrate AI into disease surveillance, resource allocation, and community health outreach without requiring deep technical expertise.
Sources: TechTarget / Health Tech Analytics Share ↗ ✉︎ Email 💬 Text
'Headless AI' aims to eliminate healthcare workflow bottlenecks
Aditya Bansod, co-founder and president of Luma Health, is championing a concept he calls "headless AI" — artificial intelligence that operates quietly behind the scenes to eliminate workflow bottlenecks without demanding clinician attention. Unlike chatbot-style AI that requires user prompts, headless AI monitors processes, detects friction, and acts autonomously: rescheduling appointments, routing messages, or flagging care coordination gaps. Bansod argues this approach is better suited to clinical environments where staff attention is already stretched thin. The model prioritizes outcome over interface, letting AI shoulder administrative burden so clinicians can focus on patients rather than systems.
Sources: Healthcare IT News Share ↗ ✉︎ Email 💬 Text
New AI risks: 'Cognitive spoofing' and fake clinical expertise
Dr. David Kirk, Chief Medical Officer at Regard, has identified a new category of AI risk he calls "cognitive spoofing" — where AI systems project the appearance of clinical expertise without possessing real medical experience or judgment. Unlike hallucinations that produce factually wrong outputs, cognitive spoofing occurs when an AI's confident, fluent responses are mistaken for clinical authority. Kirk warns clinicians may over-trust AI recommendations in high-stakes situations precisely because outputs sound credible. He calls for healthcare AI vendors to be explicit about their models' clinical reasoning limits and advocates for clearer disclosure standards before AI tools reach patient-facing or diagnostic contexts.
Sources: Healthcare IT News Share ↗ ✉︎ Email 💬 Text
Why healthcare AI systems fail after deployment
Tharun Kumar Nallamothu, Senior Software Engineer at Microsoft Health AI, identifies the real culprit behind post-deployment AI failures: not model performance, but operational integration. Writing in Health IT Answers, Nallamothu pinpoints four recurring failure points — workflow disruption, interoperability gaps across fragmented EHR environments, HIPAA compliance challenges, and latency requirements real-time clinical tools cannot always meet. "Success depends on deep embedding into clinician environments, with strong identity, compliance, and latency-aware design," he writes. "These systems only scale when reliability and governance are built in from the start, not layered on later."
Sources: Health IT Answers Share ↗ ✉︎ Email 💬 Text
A big challenge for healthcare AI: Simplify the stack
Healthcare's AI ambitions are running into a stubborn obstacle: infrastructure complexity. Patrice Bordron, Chief Digital and Information Officer at Community Health Systems, frames it plainly: "Healthcare's challenge is no longer digitization — it is simplification." Years of layered EHR implementations, point solutions, and siloed data stores have created technical stacks even experienced engineers struggle to navigate. AI tools layered on top risk amplifying existing fragmentation rather than solving it. Bordron and other health system CIOs are calling for architectural consolidation before adding more AI, arguing that simplifying the underlying stack is a prerequisite for AI that actually scales and delivers consistent clinical value.
Sources: Healthcare IT News Share ↗ ✉︎ Email 💬 Text
New genetic programming platform could advance precision medicine
Researchers at Hebrew University have developed a modular synthetic gene circuit platform that could expand the toolkit for precision medicine, according to a study published July 16, 2026 in Nature. The platform allows scientists to design and assemble gene circuits with previously unavailable modularity, making it easier to engineer cells for specific therapeutic functions. As proof of concept, the team demonstrated production of IL-15, a protein with known cancer-fighting properties. While clinical applications remain years away, the platform's modular design could accelerate drug development timelines and enable more targeted therapies for cancer and other complex diseases.
Sources: Healthcare IT News Share ↗ ✉︎ Email 💬 Text
Curated by JD · samwise.agency