Healthcare AI News 10/29/25

News

Hospital for Special Surgery hip and knee replacement patients give high marks to an AI-powered chatbot that was trained on the hospital’s patient education materials to answer their pre- and post-op questions. The small study found that most questions were asked before the operation rather than after, and patients said they were comforted by knowing that someone was always available to respond. The technology was provided by customer care AI agent vendor Aidify.

A Johns Hopkins study finds that physicians perceive their AI-using peers as less capable, regardless of whether those doctors use it for primary decisions or for verification. Doctors viewed peers most favorably when they avoided generative AI altogether, even though most said they appreciate its healthcare potential.

Cleveland Clinic expands its use of Bayesian Health’s sepsis detection software, which applies AI to EHR data to identify at-risk patients. The Clinic is an investor in the company.

The American Medical Association asks ASTP/ONC to harmonize federal AI regulations, remove regulatory barriers, and ensure that clinicians review algorithms that affect patient privacy and safety.

Business

Laudio enhances its leader operations platform with Performance Insights, an AI tool that provides insight for mentorship and performance management.

England-based Aide Health, which develops apps to help patients manage chronic conditions, launches an AI tool that records and summarizes medical visits for patients and their families. The company says that its Mirror app improves prescription adherence and engagement, reduces avoidable visits, and helps patients retain information that they would otherwise forget, which it estimates is 80% of what is discussed during appointments.

Research

OpenAI finds that 0.07% of ChatGPT users who are active in a given week exhibit possible signs of mental health emergencies in their AI conversations. The company created a network of 170 psychiatrists, psychologists, and primary care physicians to devise ChatGPT responses that encourage users to seek real-world help. The company acknowledges that while the percentage is tiny, it still represents hundreds of thousands of users.

Other

A man’s brother-in-law dies of a heart attack during a brief ED visit that generated a hospital bill for $195,000. The brother used ChatGPT to negotiate it down to $37,000 by requesting an itemized bill with CPT codes, comparing the charges to Medicare rates, and then finding major discrepancies. The hospital agreed to correct its charges, but asked him to accept the bill reduction as charity care. He declined, saying that the move was likely to protect the hospital’s tax-exempt status. He concludes that “hospitals know they are the criminals they are” and that no one should pay more out of pocket than Medicare would.

An 1,800-attorney law firm apologizes to a judge in a hospital bankruptcy case for submitting an AI-generated filing that contained inaccurate and fabricated legal citations. The hospital alerted the court that the document appeared to be AI-generated, but the attorney initially denied it, later admitting that while she hadn’t used AI herself, she knew it had been used but let it pass because she was overworked.

Contacts

Mr. H, Lorre, Jenn, Dr. Jayne.
Get HIStalk updates.
Send news or rumors.
Follow on X, Bluesky, and LinkedIn.
Sponsorship information.
Contact us.