Recall’s security issues might not have been completely eliminated.
Microsoft Recall’s security woes have come back to the fore after a test caught the AI screenshotting tool capturing sensitive data (again). Ahead of its public beta release in April, Microsoft made a slew of security updates to Recall, including adding a filter that’s supposed to block Recall from recording passwords, credit card info, social security numbers, and similar sensitive data. It looks like more fine-tuning is still needed.
The Register’s Avram Piltch conducted an in-depth security test on Recall that revealed the AI doesn’t always filter out sensitive data. The filter was usually successful when keywords like « password » or « pay » were on screen, but whenever they weren’t, Recall often misfired and took a screenshot. For instance, it screenshotted a text document with a list of usernames and passwords that weren’t labeled.
