AI inferences will cause most privacy incidents by 2029

Most privacy incidents will arise from AI-generated inferences about individuals by 2029, not from direct exposure of personally identifiable information (PII), a Gartner report said on Friday.

Advances in generative AI and machine learning let attackers extract sensitive attributes such as health conditions or behavioural patterns from anonymised or aggregated data. Bart Willemsen, VP Analyst at Gartner, said there is a fundamental shift from data exposure to insight exposure. Inference attacks often evade conventional detection mechanisms.

The report predicted spending on data integrity protections will reach parity with data confidentiality investments by 2028. Gartner urged security leaders to embed AI governance into privacy programmes, adopt privacy-enhancing technologies such as differential privacy and synthetic data, and strengthen data minimisation controls.