A cache of internal documents, screenshots, and audio recordings reviewed by Motherboard revealed that contractors working for Microsoft are listening to personal conversations of Skype users conducted through the app's translation service. These conversations include people talking intimately to loved ones, some chatting about personal issues such as their weight loss, relationship problems and even phone sex. Other files obtained by Motherboard show that Microsoft contractors are also listening to voice commands that users speak to Cortana, the company's voice assistant.
"The fact that I can even share some of this with you shows how lax things are in terms of protecting user data," a Microsoft contractor who provided the cache of files to Motherboard said on the condition of anonymity, as the person is under a non-disclosure agreement with the company.
The audio clips obtained by Motherboard are typically short, lasting between five and ten seconds, yet the source said other passages can be longer. Some of the audio obtained by Motherboard is specified as coming from the Translator feature of Skype's Android app, according to accompanying screenshots of the contractor's screen.
An FAQ for Skype Translator says that when people use the service, "Skype collects and uses your conversation to help improve Microsoft products and services. To help the translation and speech-recognition technology learn and grow, sentences and automatic transcripts are analyzed and any corrections are entered into our system, to build more performant services." Another section adds, "To help the technology learn and grow, we verify the automatic translations and feed any corrections back into the system, to build more performant services." No section of this FAQ says that the audio could be listened to by any other human person.
A Microsoft spokesperson told Motherboard in an emailed statement that "Microsoft collects voice data to provide and improve voice-enabled services like search, voice commands, dictation or translation services.” The spokesperson said that Microsoft gets customers’ permission before collecting and using their voice data and that the company takes steps to “prioritize users’ privacy before sharing this data with our vendors, including de-identifying data, requiring non-disclosure agreements with vendors and their employees, and requiring that vendors meet the high privacy standards set out in European law.”
According to the screenshots and other documents describing the official procedure, a contractor presented by Microsoft with a piece of audio to transcribe is given a series of approximate translations generated by Skype's translation system. The contractor then needs to select the most accurate translation or provide their own, and the audio is treated as confidential Microsoft information, the screenshots show. The company also said audio data is only available to contractors through a secure online portal.
"Some stuff I've heard could clearly be described as phone sex. I've heard people entering full addresses in Cortana commands or asking Cortana to provide search returns on pornography queries. While I don't know exactly what one could do with this information, it seems odd to me that it isn't being handled in a more controlled environment," the contractor said.
Despite the sensitivity of the information, it is at least in part work-at-home contractors who are listening to and handling the Skype and Cortana audio. Motherboard found online job listings from Microsoft contractors that say employees can work from home.
"I generally feel like that while we do not have access to user identifiable information, that if Microsoft users were aware that random people sitting at home in their pajamas who could be joking online with friends about the stuff they just heard that they wouldn't like that."
Earlier both Apple and Google suspended their grading programme for audio assistants as they failed to tackle security flaws, issuing an update several months after they were discovered by Google's Project. According to reports published in The Guardian, Apple's contractors all over the world regularly receive Siri recordings with users' private information without disclosing it in privacy documentation. The data includes audio requests, drug deals, medical questions, and recordings of couples having sex.
