Are social media companies responsible for the lies their users tell? Both the obvious answers, “yes” and “no”, are clearly wrong. Complete responsibility is a bad idea, and impossible in practice: even in China, the home of the largest and most sophisticated censorship apparatus on the web, the internet is expected to slow down markedly in the coming weeks under the burden of combing through it to ensure that no references to the 30th anniversary of the Tiananmen Square massacre are published. And, as the Chinese example shows, there are also considerable difficulties that arise when any one organisation can decide what counts as truth or falsehood. Yet it can’t be right, either, to say that social media companies have no responsibility to exercise the powers they have to remove obnoxious material from their servers. Videos of murder, child abuse and other horrors are routinely and rightly removed. It will be objected that these are horrible precisely because they are not lies – they record things that really happened. But that doesn’t stand up. It is no defence, either in British law, or in any moral sense, to say that a video of atrocity is faked. If it works as propaganda for jihadis, or for child abusers, it will be censored and its originators punished if that’s possible.
The platforms have been much more reluctant to act against lies which promote causes which are not in themselves criminal, however despicable. Google and Facebook are both advertising businesses, the biggest that the world has ever seen, and they depend on their ability to attract and to retain viewers. So the conspiracy theories of Alex Jones were tolerated for many years. So were the 9/11 truthers, the anti-vaxxers, and, on Twitter, Donald Trump’s campaign to suggest that Obama had been born abroad.
The sophisticated attacks of “deepfakes”, where video is manipulated to show real people doing or saying things which in real life they never did, are frightening enough. But consider what can be done with very shallow fakes indeed: last week a video was circulated on both Facebook and YouTube purporting to show Nancy Pelosi, the Democratic leader of the House of Representatives, giving a speech. It had simply been slowed down by a quarter, and then had the pitch of her voice boosted back up to a normal tone. The result was to make her appear drunk or seriously unwell. YouTube took it swiftly down. Facebook, where it has been watched at least 2.8m times, has refused to do so. Instead, it placed a sidebar on the video, suggesting as “further viewing” various news stories pointing out that it is a lie.
This will of course have no effect on those who want to believe. It looks like a deliberate business decision by Facebook to position itself as the preferred channel for dirty politics in the pursuit of power without responsibility. That is why it is the task of democratic politicians to compel the company to take responsibility for its coarsening of political debate.
