Johnson Sends Letter to Heads of Instagram, Facebook on Suggested Content Algorithms

WASHINGTON — U.S. Sen. Ron Johnson (R-Wis.), a member of the Senate Commerce Committee, sent a letter to Head of Instagram Adam Mosseri and Facebook CEO Mark Zuckerberg on Monday, requesting information on how algorithms and artificial intelligence suggest content and users to engage with on their platforms.

“After conducting a number of algorithm-related studies, Pew Research Center concluded, ‘Americans express broad concerns over the fairness and effectiveness of computer programs making important decisions in people’s lives.’ As we become aware of the society-wide significance of this influence, the lack of transparency regarding human bias and the use of algorithms and artificial intelligence is troubling,” Sen. Johnson wrote. “As a result, policymakers and the American public deserve to understand the facts behind the content and suggestions they are served on these internet platforms.”

The full text of the letter to Mr. Mosseri and Mr. Zuckerberg is below and can be found here.

The senator’s letter comes the week after he raised questions about Instagram’s follow-suggestion algorithm at a Commerce Committee hearing on social media algorithms.

July 1, 2019

 

Adam Mosseri                                                             Mark Zuckerberg                               

Head of Instagram                                                      Chairman and Chief Executive Officer

Instagram                                                                    Facebook

200 Jefferson Drive                                                    1 Hacker Way
Menlo Park, CA 94025                                               Menlo Park, CA 94025                                  

 

Dear Mr. Mosseri and Mr. Zuckerberg:

The United States Senate Committee on Commerce, Science, and Transportation recently held a hearing about the use of persuasive technology on internet platforms.  Discussed at the hearing was how algorithms and artificial intelligence (AI) are used on internet platforms such as Facebook, Instagram, Twitter, Google, YouTube and others.  These platforms have transformed the way people communicate and have the power to influence, and in the words of two witnesses, manipulate those using their platforms.  There are more than 2 billion social media users worldwide, and in 2018, there were approximately 243.6 million users in the United States alone.[1]

At the hearing, Tristan Harris, co-founder and executive director of the Center for Humane Technology, testified, “What we’re experiencing with technology is an increasing asymmetry of power that’s been masquerading itself as a[n] equal or contractual relationship.”  He went on to say, “You don’t want a system of asymmetric power that is designed to manipulate people, and you’re always going to have that insofar as the business model is one of manipulation. … I think what we need is a mass public awareness campaign so people know what’s going on.”[2]  In other words, people need to know they are being manipulated.  Given that these internet platforms have hundreds of millions of users and the asymmetric power to manipulate, a number of very legitimate questions and concerns must be raised.

The Federal Communications Commission (FCC) is tasked with regulating ownership of broadcast stations.[3]  It sets limits on television and radio ownership to ensure that one individual or entity cannot exert inappropriate levels of influence over the public.  The Department of Justice (DOJ) and the Federal Trade Commission (FTC) are tasked with enforcing antitrust laws and other regulations.  Congress conducts relevant oversight to inform public policy.  These entities of the federal government should be concerned about how much power to influence the public major internet platforms have acquired, and whether enforcement action or new laws and regulations are required.

In my questioning at the hearing, I provided the results of an internet search conducted by a member of my staff — a conservative woman in her late 20s — on her personal Instagram account.  Using her Instagram mobile phone app, she searched for Politico, clicked to pull up Politico’s page, and then hit “follow.”  She was provided “Suggestions for You” of other accounts to follow.  The suggested follows were not what one would expect based on Instagram’s Data Policy, which states:

We use the information we have to deliver our Products, including to personalize features and content (including your News Feed, Instagram Feed, Instagram Stories and ads) and make suggestions for you (such as groups or events you may be interested in or topics you may want to follow) on and off our Products. To create personalized Products that are unique and relevant to you, we use your connections, preferences, interests and activities based on the data we collect and learn from you and others (including any data with special protections you choose to provide); how you use and interact with our Products; and the people, places, or things you're connected to and interested in on and off our Products.[4]

A video of the staffer’s screen is enclosed.  Multiple other staffers of various demographics (though all are conservative) followed the same click pattern, which yielded similar results.  The first staffer I mentioned follows more than 1,300 accounts.  She estimates that the vast majority of the accounts she follows are friends, family, and accounts related to other personal interests (fashion, dogs, sports, etc.), and only 10% to 20% of the accounts are politically related, with the majority of the politically related accounts being conservative-leaning ones.  Based on this click pattern and the conservative profile of these individuals, I’d like to understand how and why this content was generated.

After conducting a number of algorithm-related studies, Pew Research Center concluded, “Americans express broad concerns over the fairness and effectiveness of computer programs making important decisions in people’s lives.”[5]  As we become aware of the society-wide significance of this influence, the lack of transparency regarding human bias and the use of algorithms and artificial intelligence is troubling.  As a result, policymakers and the American public deserve to understand the facts behind the content and suggestions they are served on these internet platforms.

To better understand the business practices of Instagram and Facebook, particularly how content and suggestions are generated, please provide a staff briefing by July 10, 2019.  Prior to a staff briefing, please provide written responses to the following questions:

  1. Specific to the above-mentioned example, which entity generated the “suggested for you” list — Politico, Instagram, or Facebook?
    1. If it was either Instagram or Facebook, please explain how this list was generated. 
  2. B.J. Fogg wrote a book in which he discusses persuasive technology, which involves interactive computing systems designed to change people’s attitudes and behaviors.[6] Do either Facebook or Instagram deploy “persuasive technology”?
  3. In what instances do Facebook and Instagram use algorithms, AI, or a mixture of both to generate user content and suggestions?
  4. Whether written explicitly in company policy or not, do Facebook and Instagram acknowledge that instances of human-driven bias within algorithms exist on the platforms? 
  5. Do Instagram and Facebook attempt either to determine or to measure the extent to which the bias of the algorithm/AI human developer influenced the results?
  6. Do Instagram and Facebook take any action to limit or accentuate human bias in search results?

Thank you for your attention to this matter. Should you have any questions about this request, please contact Mimi Strobel at mimi_strobel@ronjohnson.senate.gov or (202) 224-5323.

Sincerely,

Ron Johnson

United States Senator

###



[1] Clement, J. “Number of Social Media Users Worldwide 2010-2021.” Statista, 29 Nov. 2018, www.statista.com/statistics/278414/number-of-worldwide-social-network-users/.

[2] Optimizing for Engagement Understanding the Use of Persuasive Technology on Internet Platforms: Hearing before the S. Comm. on Commerce, Science, and Transportation, 116th Cong. (June 25, 2019).

[3] “FCC Broadcast Ownership Rules.” Federal Communications Commission, 15 May 2019, www.fcc.gov/consumers/guides/fccs-review-broadcast-ownership-rules.

[4] “Instagram Help Center.” Data Policy | Instagram Help Center, help.instagram.com/519522125107875.

[5] Smith, Aaron. Public Attitudes Toward Computer Algorithms. Pew Research Center: Internet, Science & Tech, 16 Nov. 2018, www.pewinternet.org/2018/11/16/public-attitudes-toward-computer-algorithms/.

[6] Persuasive Technology: Using Computers to Change What We Think and Do, B.J. Fogg, 2003.