Baked-in bias

‘Once social media entities become publishers, any political bias already present will be baked in’

Open Season

“Always remember,” my pupil-master used to say, “the only law continually in force is the law of unintended consequences.” Legal practitioners are less likely to forget this, but we look on in dismay as politicians do so as a matter of routine.

Facebook recently banned a number of outspoken conservative commentators from its platform, and one equally prominent progressive religious leader. The former included the alt-right Briton Paul Joseph Watson, while the latter was none other than perennial controversialist Louis Farrakhan. An enormous contretemps ensued on both sides of the Atlantic, with Donald Trump wading into the debate.

“I am continuing to monitor the censorship of AMERICAN CITIZENS on social media platforms,” the President thundered on his favoured toilet wall, Twitter. “This is the United States of America—and we have what’s known as FREEDOM OF SPEECH! We are monitoring and watching, closely!!”

Trump then set about re-tweeting the banned figures—a notable act in itself, because Twitter has traditionally been the most censorious of social media outlets.     Facebook, meanwhile, was always somewhere in the middle, with YouTube least severe of all. It was Facebook, however, that hosted the Christchurch Shooter’s live stream, becoming an unwitting vehicle for its distribution all over the internet. Times have changed.

Falling in behind Trump, even moderate conservatives and conventional media dialled up the outrage, pointing out the extent to which social media is now exercising editorial judgment like orthodox publishers.

Another, smaller group (mainly Americans) also argued that “platform access” is a civil right, eschewing the “publisher” argument. The walled gardens of social media, they suggested, are a new public square, to which everyone has access as a matter of right. This group would mandate social media adherence to the First Amendment—a striking departure from existing US jurisprudence, which only imposes 1A on the government, not private bodies.

Both these arguments are wrong. They’re wrong, however, in interesting ways. In future, it may also become necessary to pick one and make it right by force of law, a process fraught with danger. We are confronted with a civil society mess, and the only law always in force is the law of unintended consequences.

Section 230 of the US Communications Decency Act and Article 13 of the EU Copyright Directive govern social media. Both exempt YouTube and Facebook from legal liability for their users’ posts. Twitter, for example, cannot be sued if one of its users defames another user. Importantly, neither EU nor US law mandates that such entities act in a neutral fashion. Section 230, in particular, was enacted to allow online communities to engage in reasonable, good-faith moderation without fear of undue liability for their users’ posts.

The immunity provided is not absolute. Section 230 has no effect on the notice-and-takedown regime imposed by the Digital Millennium Copyright Act (DMCA), for example. Article 13 imposes a duty on large corporates that host user-created content to take active measures to prevent copyrighted material from being uploaded without permission, under penalty of being held liable for users’ copyright infringement.

However, a US doctrine known as “fair use” permits limited use of copyright material without having first to acquire permission. This provides a defence to infringement claims, and means that DCMA isn’t particularly onerous. In the UK and EU, by contrast,  “fair dealing” only provides for specific exemptions from copyright protections, and permission has to be sought from the rights-holder in advance. Article 13 states this “shall not lead to any general monitoring obligation”, but as technology currently stands, it’s difficult to see any alternative to automated filters. And making social media liable in advance for users’ IP violations is among several classic indicators of publisher status. It goes hand-in-hand with liability for defamation.

So, YouTube and Facebook and Twitter aren’t publishers in the eye of the law—yet. That could change. Exercising editorial control in a politically-slanted way is an open invitation to legislators to amend or repeal Section 230 or the Copyright Directive or, in the UK, treat social media as broadcasters and drag them under Ofcom’s aegis.

This would delight traditional media by levelling the competitive playing field, forcing Facebook and YouTube to hire armies of lawyers and exercise editorial oversight that isn’t only political but also attuned to quality control. That’s why The Times supports it. Just as the only law always in force is the law of unintended consequences, one should always bet on self-interest, because it’s the only horse that’s trying.

Relatedly, conservatives who think the way to stop social media censoring right-wing pundits is to treat web companies in the same way as the BBC are delusional. Once social media entities become publishers, any political bias already present will be baked in. Like a newspaper or cable channel, they’ll have a slant. Twitter will be left, Facebook and Instagram centre-left, YouTube centre-right to right, and so on. It would mean much less opportunity for untrammelled speech of any sort.

One thing publishers do (and have always done) is act as gatekeepers. Not everyone has a column in Standpoint. Even getting a letter to the editor into the newspaper or a right-of-reply on the BBC is dependent on appealing to gatekeepers. The only way to impose neutrality on social media would be to treat it as a utility, like Royal Mail or BT. Utilities are not liable for any content but are also not able to engage in censorship (or moderation) except as required by law. Royal Mail can’t refuse to deliver your post unless you use it to commit a crime. And despite the hopes of some Americans, treating social media as a utility is vanishingly unlikely in the UK or EU and even less likely in the US, given the way the law there is framed.

It’s possible this vast experiment where everyone could speak at once will soon come to an abrupt halt. I suggest—before we return to how things once were—that we undertake an honest appraisal of where we are now.