Transparency for Internet Companies’ Political Filtration of User-Generated Content Prepared written testimony of Prof Eugene Kontorovich Professor of Law George Mason University Antonin Scalia School of Law Hearing before the Senate Judiciary Committee Subcommittee on the Constitution Stifling Free Speech Technological Censorship and the Public Discourse April 10 2019 1 Chairman Cruz Ranking Member Hirono and honorable members of the Committee thank you for having me here to testify on constitutional issues relating to the alleged ideological censorship of online content by large internet platforms I am a professor at George Mason University Antonin Scalia Law School having previously taught for more than a decade at Northwestern University School of Law I teach constitutional law and have written extensively on First Amendment issues in a variety of contexts I should begin by stating that the extent to which internet platforms engage in politically biased content-sorting or are more particularly biased against conservatives is a factual question on which I can claim no expertise I will assume for the purposes of the present legal analysis that such an issue exists to some extent and consider what approaches Congress might take in responding I It is not a First Amendment problem The first thing to note about the subject is that it is not a First Amendment issue The First Amendment only applies to censorship by the government indeed the text of the Amendment speaks only of the federal government The conduct of private actors is entirely outside the scope of the First Amendment If anything ideological content restrictions are editorial decisions that would be protected by the First Amendment Nor can one say that the alleged actions of the large tech companies implicate “First Amendment values ” or inhibits the marketplace of ideas in ways analogous to those the First Amendment seeks to protect against The First Amendment – like the Due Process Clause does not have a penumbra of values beyond what its text prohibits Thus the “values” behind the First Amendment are nothing more than prohibiting governmental restraints on speech This point is often obfuscated in a variety of contexts for example some supporters of net neutrality have incorrectly claimed that it is necessary to protect free speech values While companies like Facebook and Google serve as conduits of information for vast numbers of people and have significant market power their power and permanence cannot be compared to that of the government Such companies position in the market may seem unassailable now one should recall that similar arguments were made about Microsoft’s Internet Explorer by its then rival Netscape The mere fact that ideologically-biased content moderation by tech companies does not raise anything like constitutional “free speech” issues does not mean it is not a legitimate matter of public concern and discussion It is certainly reasonable for Congress to take ideologically biased practices by purportedly neutral internet firms into account when considering updating or revising existing legislation In particular the marketplace of ideas requires participants to know what the rules are If information is being filtered or blocked this is the First Amendment right of the internet companies but Congress can also consistent with the First Amendment require that customers – those using the online services – be informed of such censorship Moreover existing law already privileges online companies in ways not required by the First Amendment The remainder of my testimony will discuss these points 2 II Section 230 immunity Current regulation is already not neutral on the subject of internet content providers The Communications Decency Act of 1996 provides special protections for “interactive computer services ” Such companies cannot be treated as publishers of content on their sites created by third-parties even if they would otherwise qualify as publishers under common law 1 This provision shields companies from liability for much of the content on their sites on the theory that they are not acting analogously to newspaper editors – making substantive decisions about what appears and what does not – but simply providing a forum or platform The law also provides internet companies with a very broad immunity to civil liability for taking down or blocking “objectionable” material These statutory provisions have likely had a helpful role in the growth of the Internet But they are also clearly not constitutionally required – they are special favors If indeed these online services’ role in moderating content is sufficiently minimal that they should not be regarded as publishers that is presumably something they could prove in court under preexisting legal standards on an equal footing with “Old Media ” Section 230’s blanket presumption is not mandatory and certainly open to revision in a different environment more than two decades after its passage In enacting the immunity provisions Congress assumed that protected internet services provide “a forum for a true diversity of political discourse ”2 To the extent that assumption is weakened by online companies filtering out viewpoints that they deem ideologically impermissible the assumptions behind Section 230 may need to be revisited When sites block content expressing certain extreme ideologies but not other extreme ideologies they are arguably making editorial choices that make them seem more like publishers Many have argued that any change to Section 230’s exemptions from liability will result in more filtering and blocking of potentially controversial content That may well be true and is certainly a policy question this body should consider But such greater filtering will likely not have an ideological tilt – civil liability is decidedly non-partisan And it is a policy question rather than a constitutional one whether it is better to have greater across-the-board content screening or less screening that is more politically unbalanced III Consumer transparency and disclosure Politically biased filtering and blocking practices by internet companies raise legitimate questions about consumer protection and transparency Again I will stress that companies like Facebook and Twitter are free to remove whatever content they like The “marketplace of ideas” is strong enough to withstand this – if conservatives find these companies allegedly biased content policies objectionable they can migrate to or create other platforms To be sure there are switching costs with network goods but the Internet also allows for the rapid creation of new networks existing networks such as Facebook and Twitter also compete against each other However such a market response depends on consumers being aware what is happening 1 2 47 U S Code § 230 c 47 U S Code § 230 a 3 3 One of the greatest strengths and attractions of platforms like Facebook Twitter and Google is the expectation of users that through the window of their computer they can see or access the vast parade of information of the world Users understand the experience of “following” as their selection from amongst the myriad accounts and profiles of the platform which they will follow Their feed then gives them a perspective on outside events This perspective is of course filtered by their own preferences and predilections in selecting what to follow which many social critics lament creates self-reinforcing “information silos ” confirming users’ prior beliefs and prejudices This problem is well-known and natural – no different from people subscribing only to magazines that reflect rather than challenge their beliefs Consumers can anticipate this problem if it is a problem for them because it is a result of their own online choices What users may not anticipate is that the online marketplace of ideas has already been ideologically pre-sifted by the online platform in a way that is not neutral Users that turn to a search engine to discover things about the external world will have a poor understanding – we all have a poor understanding – of how the search results may have been manipulated on politically non-neutral criteria They will assume that ideas they do not see do not exist This problem is exacerbated by the increasing tendency of journalists to turn to social media as representing a cross-section of public opinion the man-in-the-street interview has been replaced with the tweet-on-the-feed The prior filtration of social media content by the platform changes the way people understand the world without their necessarily knowing it A recent example will illustrate this A feature film called Unplanned currently playing in theaters takes a position critical of abortion The film has naturally been controversial and its reception by various arbiters of standards has allegedly been biased Thus the film-makers complained that the decision of the Motion Picture Association of American to give it an “R” rating was politically motivated – that the rating was unduly restrictive given the movie’s content Then on the film’s opening weekend its Twitter account was suspended and tens of thousands of followers inexplicably dropped The MPAA action regardless of whether it was politically neutral was transparent consumers and reviewers could see the rating and compare it with what they knew of ratings for other films A disappearing social media account or followers is by definition not transparent one can’t see what is not there When users of Twitter found out about these developments from press accounts it resulted in significant online protest and the restoration of the account The processes and policies by which internet companies filter block and rank content is entirely obscure to consumers The extent to which such companies have policies that disfavor conservative content is a factual issue on which I can take no position However if Congress concludes such a problem exists one obvious solution consistent with both free speech and free markets is to require disclosure to consumers of policies or algorithms that sort content on substantive politically biased grounds 4 A disclosure requirement need not be onerous and would be far less intrusive than any revision of Section 230 In a related content a similar “transparency” principle was adopted by the Federal Communications Commission for ISPs in lieu of the net neutrality rule 3 I thank the Committee for its time and will be happy to answer any questions 3 See Federal Communications Commission Restoring Internet Freedom Order WC Docket No 17-108 FCC 17-166 Jan 4 2018 47 CFR § 8 1 ISP consumer transparency regulations 5