Researchers develop a noise-exploitation attack which could defeat dynamic anonymity

According to TechCrunch, a group of privacy researchers argues they have gathered the first evidence that a long-theorised vulnerability in systems anonymising data by aggregating noise, is ‘no longer just a theory’. These findings add up the growing body of research suggesting that the risk of de-anonymisation on large datasets is persistent. This work was presented at the last USENIX Security Symposium by a team of researchers from Imperial College London and Université Catholique de Louvain.