Healthcare AI News 2/25/26
News
Anthropic introduces workflow plugins for Claude Cowork that allow users to connect to enterprise software, develop private plugin marketplaces, and deploy AI agents. New connectors include Google Workspace, Docusign, WordPress. New plugins support HR, design, operations, brand voice, financial analysis, and equity research.
An NVIDIA survey of 600 healthcare and life sciences executives and AI practitioners finds that 70% of organizations are actively using AI, with medical technology and drug companies reporting return on investment and nearly half deploying agentic AI. The most common use case is data analytics and data science. Among management respondents, 85% say that AI has increased revenue, and 80% say that it has reduced costs.
Business

Anthropic announces that its Claude Code tool can modernize COBOL programs, sending IBM shares down 13% in their biggest one-day drop since 2000 and wiping out $40 billion in market value. IBM responds that COBOL modernization has been a solved problem for years and that the real issue is cost and return on investment. Analysts say that Anthropic could take some market share from IBM’s tooling, which they believe provides minimal revenue. IBM shares have rebounded slightly, but have lost 9% in the past 12 months.

Publicly traded business software vendor UiPath announces agentic AI solutions for healthcare that include medical record summarization, claims denial prevention and resolution, and prior authorization support.

B. well Connected Health launches a white-label health AI assistant that it says can be deployed by app developers in just a few weeks.

Ardent Health will deploy Hellocare.ai’s AI-assisted virtual clinician and patient observation platform in 2,000 patient rooms.
Research
A study finds that ChatGPT Health failed to recommend emergency care in more than half of serious cases compared to physicians. In some instances, the tool appeared to recognize a serious condition but still offered reassuring guidance, prompting researchers to conclude that a disconnect exists between its clinical understanding and its recommendations. The authors also report that the tool performed inconsistently in directing users with suicide risk to a crisis hotline.
Other

A hospital and medical school in Thailand deploy AI-powered robots to support the care of patients with thyroid cancer and tuberculosis.
Contacts
Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Follow on X, Bluesky, and LinkedIn.
Sponsorship information.
Contact us.
![]()

Weird that Google can acknowledge its crowdsourced medical advice was wrong, but escape penalties for doing it wantonly at scale.