YouTube, Facebook, and the responsibilities of intermediary gatekeepers

YouTube logo, via YouTubeIn my previous post, I argued that, as a matter of principle, the controversial American anti-Islamic video should not be censored. The most obvious form of censorship comes from government action, such as legislation banning speech, but that does not arise in this case. Less obvious, but no less insidious, was the White House request to Google to re-consider whether the video breached YouTube rules. This was not a formal ban, and Google declined to take the video down in the US, but it did block access to it in in Egypt and Libya. This raises two important questions about the structure of free speech. First, in the online world, where most of us access the internet through a range of intermediaries, government censorship does not necessarily need to target the disfavoured speech; it need only target the intermediaries. Very few US companies would feel able to decline a request like that from the White House, and Google are to be commended for standing firm in those circumstances. Second, these intermediaries now have a great deal of practical power over online expression – not only can they be co-opted by government as agents of state censorship, but they also have the capacity to act as censors in their own rights, as Google did in their unilateral action to block access in the Middle East.

Such intermediaries are effectively gatekeepers are those who enable – and control – our access to that information, and this raises profound issues of principle about the role of intermediary gatekeepers in the structure of free speech about which I have written on this blog (here | here | here). At present, such intermediary gatekeepers are all private entities, operating to their own rules, and it is not at all clear how they can be made accountable to their users or the wider public for their private actions. Given the practical, social and legal issues that arise in policing content in such a quasi-public sphere (pdf), it has been argued that search engines and other intermediaries should have public interest obligations, perhaps by analogy with common law duties that govern public utilities (pdf). In particular, free speech norms should not only be about protecting speakers against a heavy-handed state but also about protecting speakers and readers against heavy-handed intermediate gatekeepers. This debate is now being played out online and on the op-ed pages of US news papers. Take, for example, the Washington Post:

Google’s restricting of anti-Muslim video shows role of Web firms as free-speech arbiters

Google lists eight reasons on its “YouTube Community Guidelines” page [here] for why it might take down a video. Inciting riots is not among them. But after the White House warned Tuesday that a crude anti-Muslim movie trailer had sparked lethal violence in the Middle East, Google acted.

… Legal experts and civil libertarians, meanwhile, said the controversy highlighted how Internet companies, most based in the United States, have become global arbiters of free speech, weighing complex issues that traditionally are the province of courts, judges, and occasionally, international treaty.

The legal experts quoted in the article include Tim Wu (whose masterful book, The Master Switch, examines just these kinds of issues), Andrew McLaughlin (formerly of Google and the White House), Jennifer Granick (Director of Civil Liberties at the Center for Internet and Society at Stanford Law School) and Jonathan Zittrain (whose prescient book, The Future of the Internet — And How to Stop It, examines just these kinds of issues). Their collective point is this: intermediary gatekeepers represent a new potential for censorship, both by states acting on them, and by them acting on their own initiative; legal principles have evolved to cope with the former; they need now to evolve to cope with the latter.

Facebook logo, via FacebookThese issues arises many many times every day. By of further example, more humourously, but perhaps even more seriously, a similar spat occurred between The New Yorker magazine and Facebook. Both YouTube and Facebook require membership before data can be posted on the site, but YouTube can be searched by anyone, whereas access to most Facebook content is restricted to members. In many ways, this affords Facebook even greater scope for censorship than YouTube: not only can it control what is posted in the first place, it can also control what its users get to see. The facts of The New Yorker spat with Facebook are funny in all sorts of ways, but the issue relating to the structure of free speech is profound:


… we got temporarily banned from Facebook for violating their community standards on “Nudity and Sex,” by posting [a] Mick Stevens cartoon [featuring Adam and Eve, with dots for their nipples] … while female nipple bulging … is a potentially serious problem, with as yet no known cure, it also has no known victims. That is, unless you count freedom of expression, common sense, and humor.

You’ll have to head over their story to see the cartoon, but it is very innocent (as well as clever and funny). As Paul Alan Levy comments, this takedown “stands as a stark reminder of the fact that when so many members of the public have made Facebook “a key platform for expression,” we should remember how easily the arbitrary application of guidelines by relatively low-level staff can cut off important expression”. As Andrew Sellars points out, writing about the structural weakness of internet speech, intermediary gatekeepers have set the bar for free speech lower than America’s First Amendment, without much by way of a formal remedy against them where users feel aggrieved about access bans, take-downs, and other forms of censorship.

If intermediary gatekeepers are able to rely on free speech norms to resist government censorship, then they should respect exactly the same norms vis-a-vis their users. Until regulation exists to ensure this, the best way to achieve this end might well be to allow the users to rely directly upon constitutional rights as against the intermediary gatekeepers. Ironically, whilst this may be increasingly possible in European legal orders (see O’Cinneide (2003) 3(1) HLJ 77), but perhaps less so in the US, where the absence of state action on the part of the gatekeepers would preclude such a claim. It would be a fitting recognition of the importance of free speech online if the regulation of gatekeepers proved to be the context in which the US courts modified the state action bar for the emerging quasi-public sphere.