Microsoft contractors are manually reviewing voice recordings gathered via Skype’s automated translation feature and the Cortana voice assistant, a new report from Motherboard has revealed. In audio recordings shared with the publication, users could be heard having intimate conversations and discussing relationship issues as well as other personal topics like weight loss. “The fact that I can even share some of this with you shows how lax things are in terms of protecting user data,” the contractor who shared the audio files with Motherboard said.
The unnamed contractor describes having heard what they describe as “phone sex” as part of the review process, and they also said that they have heard users entering their full addresses using Cortana commands as well as using the voice assistant to search for pornography.
“The fact that I can even share some of this with you shows how lax things are in terms of protecting user data.”
The contractor pointed out that they didn’t have access to user-identifiable information, but Microsoft customers probably wouldn’t like the fact that their private conversations were being joked about by “random people sitting at home in their pajamas.” Motherboard has since found job listings that confirm that contractors are able to work from home. A spokesperson from Microsoft said that all of its vendors agree to confidentiality agreements, and that Microsoft has audit rights to ensure compliance.
The revelations come after Apple, Amazon, and Google have faced intense scrutiny over their handling of voice data, obtained over the process of operating their respective voice assistants. In light of the revelations, Apple has said that it’s temporarily paused its practice of letting its contractors listen to recordings, while Google has suspended its own program in Europe while regulators investigate the practice. Meanwhile, Amazon now explicitly discloses that humans will manually review voice data, and it gives them the ability to opt out.
The purpose of having contractors listen to recordings is to improve the voice services they’re used with. In Microsoft’s case, Motherboard reports that some contractors are tasked with listening to audio from Skype’s machine translation service, which launched back in 2015. After manually translating each clip, they are presented with a number of machine-generated translations, from which they pick the most accurate.
“Sentences and automatic transcripts are analyzed and any corrections are entered into our system, to build more performant services.”
Microsoft does disclose in both its Skype Translator FAQ and Cortana documentation that it uses voice data to improve its services, noting that “sentences and automatic transcripts are analyzed and any corrections are entered into our system, to build more performant services.” Elsewhere, it says that it “[verifies the] automatic translations and feeds any corrections back into the system.” However, Motherboard notes that it does not explicitly say that a human may listen to these recordings.
When contacted for comment, Microsoft said it only collects and uses voice data on an opt-in basis, and pointed us towards the voice section of its privacy dashboard, which allows you to delete the voice data that the company holds on you.
In a statement a spokesperson from Microsoft said that the company strives “to be transparent about our collection and use of voice data to ensure customers can make informed choices about when and how their voice data is used,” adding that “Microsoft gets customers’ permission before collecting and using their voice data.” They explained that this voice data is used “to provide and improve voice-enabled services like search, voice commands, dictation or translation services.”
“We also put in place several procedures designed 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,” the spokesperson said, “We continue to review the way we handle voice data to ensure we make options as clear as possible to customers and provide strong privacy protections.”
Update August 7th, 1PM ET: Updated with additional details from Microsoft.