'Think before you type': Expert explains how Yahalomi case shows dangers of sharing all with an AI

Cyber expert Osher Cohen warns that the Yahalomi case, in which two Israelis were tracked via their AI chat history, demonstrates the privacy risks of oversharing with chatbots. He advises treating AI as a cloud service, not a private confidant, and avoiding inputting sensitive data like passwords or credit card details. Cohen emphasizes a behavioral approach: check privacy settings, delete history, and create internal AI policies for businesses.

Cyber expert Osher Cohen warns that the Yahalomi case, in which two Israelis were tracked from Vienna to Buenos Aires using their AI chat history, highlights the privacy risks of oversharing with chatbots. Cohen, founder of Yo Secure, explains that users often treat AI as a private confidant, but it is a cloud service connected to accounts, devices, and data retention policies. He advises a new habit: 'think before you type.' Sensitive data like passwords, credit card details, and legal documents should never be entered. The safest approach is behavioral, not just technical—users should check privacy settings, delete history, and avoid uploading sensitive documents. Businesses should create internal AI policies. Cohen concludes: 'The danger is not only what AI knows how to answer. The danger is what we choose to tell it.'