Microsoft Updates Privacy Policy to Acknowledge Human Review of Audio
Microsoft revises its privacy policy to explicitly state that human workers review select Skype and Cortana audio snippets to improve AI accuracy.
Microsoft updates its privacy policy to explicitly confirm that human workers listen to select Skype and Cortana audio recordings. The revised policy states that the company uses both automated and manual methods to process personal data, relying on people to review short, de-identified voice snippets to improve speech recognition and translation services.
This policy change follows a Motherboard report revealing that contractors regularly listen to personal Skype conversations and Cortana voice assistant audio. Microsoft admits it made the update because it realizes it needs to do a better job of specifying that humans sometimes review this content after recent questions arise about its data practices.
Microsoft is the latest in a string of major tech companies facing scrutiny over this hidden AI training practice. Apple, Amazon, Facebook, and Google all face similar backlash for failing to clearly disclose that human contractors access user audio, a requirement under European privacy laws that these companies previously overlook.