Peters Requests Information from Facebook on Handling of User Audio

WASHINGTON, DC – U.S. Senator Gary Peters (MI) today sent a letter to Facebook CEO Mark Zuckerberg requesting additional information on the company’s handling of personal audio accessed from users. The letter comes amid reports that Facebook not only captured audio from users, but that the audio was sent to a third-party vendor to be analyzed and transcribed. Last year, Peters questioned Zuckerberg at a joint Commerce and Judiciary Committee hearing regarding concerns that users’ audio conversations were listened to by Facebook.

“I asked you specifically [at last year’s hearing] if Facebook uses audio obtained from mobile devices to enrich personal information about its users. Your emphatic answer was no,” Peters wrote to Zuckerberg. “This week, reports disclosed that Facebook has been paying third-parties to transcribe private audio conversations of Facebook users. If these reports are accurate, I am concerned that your previous testimony before Congress appears to have been, at best, incomplete.”

In the letter, Peters noted that Zuckerberg later clarified in written form from his testimony last year that Facebook may, in certain circumstances, access audio from users who opted-in. He did not, however, disclose what is done with the audio after it is collected, the extent of its collection, and the reason for the discrepancy from his initial testimony.

You can read the full letter below, or by clicking here:

Dear Mr. Zuckerberg,

I am writing to request information regarding recent reports that, counter to your prior testimony to Congress, Facebook Inc. (Facebook), has collected, transcribed, analyzed, and used data from private audio conversations of Facebook users. This week, reports disclosed that Facebook has been paying third-parties to transcribe private audio conversations of Facebook users.

If these reports are accurate, I am concerned that your previous testimony before Congress appears to have been, at best, incomplete. On April 10, 2018, you testified before a joint Senate Judiciary and Commerce, Science, and Transportation Committee hearing on issues regarding social media privacy and uses and abuses of individuals’ data.  At that hearing, I asked you specifically if Facebook uses audio obtained from mobile devices to enrich personal information about its users. Your emphatic answer was no. Your exact words to me were: “You’re talking about this conspiracy theory that gets passed around that we listen to what’s going on on your microphone and use that for ads. We don’t do that.”

In written responses to questions for the record, you acknowledged that in certain circumstances, Facebook does in fact access users’ audio, when they have opted in and are using a particular Facebook service.  You wrote that, “Facebook does not use users’ phone’s microphone or any other method to extract audio to inform ads or to determine what they see in their News Feed [and that] Facebook only accesses users’ microphone if the user has given our app permission and if they are actively using a specific feature that requires audio (like voice messaging features).” Never in your follow-up responses did you articulate what you do with the audio accessed under those circumstances, the extent of Facebook’s use of this practice, or the reasons for the discrepancy in your testimony on this issue during the hearing.

Now, over a year later, reports state that Facebook was paying hundreds of outside contractors to transcribe audio clips obtained from users and that this practice only ended about a week ago. In the report, Facebook confirmed that it was transcribing users’ audio and said it will no longer do so. Facebook also said that the purpose was to ensure Facebook’s artificial intelligence correctly interpreted messages on its Facebook Messenger app. The reports state that Facebook users whose audio was obtained agreed to have their voice chats transcribed. However, the reports also state that Facebook has not disclosed to its users that their audio may be reviewed by third-parties. This disclosure is all coming just a few weeks after your company completed a $5 billion settlement with the Federal Trade Commission following an investigation into your privacy practices.

To better understand the exact nature, extent, and purpose of Facebook’s use of user audio recordings and its broader use of user data, I respectfully request that you provide the following documents and information as soon as possible but no later than August 28, 2019:

1. For what use and purpose does Facebook collect, transcribe, and/or analyze audio conversations of its users?

1. For which Facebook products or services, including entities and applications owned by Facebook, are audio communications collected, transcribed, or analyzed?

1. What is the process by which users opted in to have their audio transcribed for the Facebook Messenger App?

1. When users were asked to opt in to having their audio transcribed, was it made clear to them that this transcription would be done by human beings?

1. Please provide a detailed explanation of all of Facebook’s uses for the transcribed audio that Facebook has collected.

1. How does Facebook secure and safeguard audio files collected from users, including files sent to third parties and contractors, and what retention policies apply to these third parties?

1. Does Facebook or the third parties receiving Facebook’s user information have the ability to modify transcripts of audio recordings?

1. Does Facebook use audio files for ad targeting?

1. Does Facebook sell audio data or the transcripts of its users’ conversations to third parties? If yes, do those third parties use the information received for targeted advertising? Please provide a copy of any contract entered into with a third party regarding the sale of user audio data or transcripts.

1. Please provide a copy of all contracts for audio transcription.

1. Please provide all documents related to obtaining a user’s consent to have their audio transcribed.

Thank you for your prompt attention to this request.

###