Gary, Jeff, and Ashkan Soltani. “First Things First: Online Advertising Practices and Their Effects on Platform Speech.” RSS, 21 Aug. 2019, knightcolumbia.org/content/first-things-first-online-advertising-practices-and-their-effects-on-platform-speech. 

University of California, Berkeley

Authors Jeff Gary and Ashkan Soltani argue that in order to reduce the spread of misinformation and harmful content on social media, rather than mandating specific content based regulations, it is more effective to address the business models under which social media platforms operate .

According to the authors, the underlying platform business models encourage and enable the spread of harmful content, including fake news. The business practice of most of these private company platforms is mainly based on tailored and targeted advertising. 

What enables targeted advertising is data collection from users. And in order to collect more data about the interests, preferences, personality, and behavior of specific users, social media platforms need to maximize the frequency and intensity with which users users interact with their platform.

Misinformation, harmful speeches, and conspiracy theories are financially advantageous to the platforms because they increase engagement of users and illicit responses. Fake news may be shared by users when it is is reaffirming of their sense of identity, community, and culture. When users spend more time on these platforms, more of their data can be collected and the more exposed they are to targeted advertising.

It is not just users that promote divisive and controversial contents; the platforms themselves promote such speech to garner user attention. Examples that the authors provide are Facebook's permission, until recently, to use the category "Jew haters" by advertisers for targeted advertisement and how Facebook distributed anti-vaccination ads to potential young mothers.

Addressing the Problem of Social Media Misinformation by Focusing on Social Media Companies' Business Practices

According to Gary and Soltani, it would be more effective to employ a framework for regulating platforms not by mandating specific content based regulations, but by addressing the financial incentive structure created by targeted advertising model.

Specific content based regulations such as built in algorithms that detect fake news and acts of blacklisting and whitelisting don't always work and raise constitutional concerns with regards to free speech and civil rights.

Problems with blacklisting: it is usually done after huge pressure from the public; banning accounts is disadvantageous for social media platforms because their business model depends on the number of active users; accounts are banned after they violate a policy; it requires platforms to identify pseudonym accounts, which infringes on the right to free association and to speak anonymously

Problems with whitelisting: relies on determinations about the speaker, not the speech; there are many mainstream new sources and speakers that engage in the spread of misinformation, and it legitimizes and grants recognition to such outlets and speakers

Problems with direct content filtering: creates the most direct potential issues with the First Amendment; private platforms can legally remove content and have almost limitless power to shape online speech, promoting voices that are financially beneficial to them while suppressing others

What the authors propose: for congress to challenge the structural practices of the platforms; increasing data privacy and security by restricting how personal data can be collected and used for advertising; restricting the kind of data that can be gathered; ensuring that certain kinds of data don't be used for commercial purposes; restrict the kind of data that could be shared with third parties; focusing on business behavior rather than individual user behavior

Why Content Based Regulation is not the Most Effective Way to Combat Misinformation

Blacklisting refers to the banning of accounts that have violated platform policies. 



Blacklisting 

Whitelisting refers to trust based ranking and the legitimization of information sources through "verified" badges based on the identity of the speaker, rather than the speech itself.


Learn Before

Related