Medical records could be revealed by AI training-data vulnerability

Medical diagnosis AIs can be tricked into telling whose data trained them

AI models used to help diagnose medical conditions have a problem: They’re ready and willing to identify patients whose data was used to train them. German researchers reported in a Nature paper published Wednesday that discriminative AI models - those used to classify data and make predictions about new inputs based on their training sets - are particularly susceptible to membership inference attacks (MIAs) that query the models in an attempt t…

Medical records could be revealed by AI training-data vulnerability

Identification risks are more severe for underrepresented groups in the training data — plus, evidence that the Universe is more uneven than assumed. Hear the biggest stories from the world of science

Medical AI Model Privacy Risks

Research led by the Technical University of Munich shows that data from some individuals used to train medical artificial intelligence (AI) models could be at much higher risk of exposure due to cyberattack than others. Writing in Nature, the researchers explain that underrepresented groups, such as people with a rare disease or a minority ethnicity, are at particularly high risk of having their data exposed. A type of cyberattack called a “memb…