China’s Kimi AI Hit by Privacy Breach Exposing Users’ Résumé Data

Kimi, developed by Chinese startup Moonshot AI, is a generative AI model designed for processing long texts. Users often upload résumés and other documents for analysis. If backend session isolation fails, personal information such as names, phone numbers and employment histories can be exposed, highlighting the risks of entrusting sensitive data to large language models.

Kimi recently suffered a privacy incident ahead of a version update, with some users receiving strangers’ résumé details and phone numbers in response to their queries. This indicated that data may have been mistakenly transmitted between different accounts. As of the publication of the related reports, Moonshot AI had not disclosed the exact date of the incident, the number of people affected, the scope of the data exposure or the amount of losses.