It is common practice to share large sets of data for research purposes, which includes medical images such as MRI scans. These MRI scan images are required to be de-identified (name of participant, date of the scan, etc.), but the facial images still remain accessible. Using publicly available facial recognition software, a group of researchers at Mayo Clinic were able to correctly match patients with their MRI images 83% of the time. These findings bring up questions about and how to protect patients privacy and security in a time where technology has outpaced our laws. It also raises concerns about the future of public data set sharing, and how to continue to recruit volunteers for studies.
