Johns Hopkins alerts 3,182 patients after worker uploads data to unapproved AI tool

Advertisement

Johns Hopkins Bayview Medical Center in Baltimore is notifying about 3,182 patients that their information may have been compromised after a workforce member uploaded a spreadsheet containing patient data to a public AI tool the hospital had not approved. 

According to an Oct. 23 news release Johns Hopkins shared with Becker’s, the workforce member mistakenly uploaded the spreadsheet Sept. 21 while conducting an internal review, and the hospital learned of the disclosure the same day.

The spreadsheet included information related to admissions and treatment from May through September 2026. That information may have included patient names, admission and treatment dates, medical record numbers, account numbers, departments of service, medications administered and provider names. It did not include Social Security numbers, financial account information or other identifying information.

After discovering the incident, Johns Hopkins Bayview launched an investigation. The hospital had the workforce member delete the initial AI prompt, thread and document from the tool’s history. It also asked the company that owns and operates the AI tool to delete the information and implemented corrective measures to prevent similar incidents.

Johns Hopkins Bayview said it is reinforcing its policies and training on the handling of protected health information, including restrictions on unauthorized applications and AI tools and reeducation on the tools the institution has approved.

The hospital began mailing letters to affected patients Oct. 22. The letters include instructions for enrolling in 12 months of complimentary credit monitoring through Experian IdentityWorks. The hospital has also set up a toll-free call center to answer patient questions.

“Johns Hopkins is committed to maintaining the privacy and security of its patients’ information and is taking this incident very seriously,” the hospital said in the news release.

The release did not identify the AI tool involved.

Advertisement

Next Up in Artificial Intelligence

Advertisement