Computational privacy aims to allow data to be used while protecting individual privacy in the digital era. Anonymization techniques like removing identifying information and blurring indirect identifiers allow data to be analyzed without obtaining consent. However, uniqueness and inference pose challenges. Unique patterns in mobility data can often re-identify individuals, and much can be inferred about people from large-scale metadata even if anonymized. New techniques aim to find an actual trade-off between privacy and utility through secure systems that keep data localized instead of anonymization alone.