Shortly after insurrectionists stormed the US Capitol in January 2021, Twitter took away Donald Trump’s favourite megaphone. The platform permanently banned him from its service for incitement of violence, cutting him off from his 90m followers. Other services, from Facebook to PayPal, quickly echoed Twitter’s decision. This created an odd situation: one of the world’s most powerful leaders was unable to speak to audiences on the world’s most popular internet platforms.
For some conservatives, the de-platforming of Trump was proof positive that social media was yet another system rigged by the left to favour their preferred ideologies. Prominent right-wing figures, like conspiracy theorist Alex Jones, had previously been de-platformed, radically shrinking their audience. New social media platforms, like Trump’s own Truth Social, sprang up as alternatives to the “censored” mainstream, though none have found audiences comparable to the likes of Twitter and Facebook.
Another response came from the legislatures of the states of Florida and Texas. Both passed laws requiring social media companies to be ideologically neutral in their content moderation. In other words, these platforms needed to ensure that right-wing speech was not being treated differently from left-wing speech.
It is unclear whether the suspicion that right-wing speech is treated unfairly by these platforms is correct. Figures like Trump and Jones had been removed for violating the terms of service by harassing users, inciting violence and spreading misinformation, not for their ideology.
Furthermore, studies suggest the opposite might be true. In 2021, Twitter released a study of its internal content algorithm that concluded right-leaning politicians experienced more amplification than those on the left. But research like this has not changed the view of many on the right, that these platforms are biased against them.
A US federal appeals court quickly blocked the Florida law from taking effect, on the grounds that it limited the rights of platforms to choose what speech to amplify or suppress. The federal court in Texas recently came to a different conclusion. It upheld the law and ordered it into effect—which creates a number of complex situations.
First, it is unclear how platforms might remain ideologically neutral, and yet still control the enormous amount of bad behaviour that happens on them all the time. Platforms like Facebook delete millions of pieces of content every day before it is ever published because it is likely spam, coordinated disinformation or illegal content, like child sexual abuse imagery. The Texas law does not require Facebook to post illegal content, but it makes it harder for such platforms to control content sometimes referred to as “lawful but awful”. Abusive or harassing speech normally removed under terms of service might now be protected as political speech.
The Texas and Florida laws will almost certainly go to the US Supreme Court, where it is unclear how they might be decided. At issue will be a battle of analogies. Are platforms like Facebook and Twitter more like newspapers, or the telephone system? If they are like newspapers, they have a strongly protected editorial right to choose what content they allow to remain on the platform and what they choose to remove. Newspapers have first amendment protections, meaning that no one can demand a newspaper print their letter to the editor or cover an issue they feel is insufficiently reported on. Part of a newspaper’s freedom of speech is its editorial freedom to choose what perspectives to amplify and which to suppress.
By contrast, under US law, telephone companies are considered common carriers. All individuals have a right to conversation through them, so long as it is not lawbreaking. You are not protected if you are using the telephone to extort money from individuals. However, the presumption is that conversations on the telephone network are not to be monitored or filtered in any way.
Obviously, Facebook and Twitter are neither newspapers nor telephones. A user on Twitter—say Trump—might reach 90m individuals with a single tweet. That’s a lot more powerful than the telephone network, which is largely used for one-to-one conversations. This ability to reach a large audience feels analogous to the power of a newspaper. On the other hand, Facebook certainly does not carefully examine every utterance made by its two billion users to determine whether those statements are true and not defamatory, as a careful newspaper editor does.
One set of rules for social media is rarely appropriate for every situation
This problem might be best solved by legislating “rules of the road” for online spaces. The US last took a significant pass at those rules in 1996, with Section 230 of the Communications Decency Act. Section 230, sometimes referred to as “the 26 words that created the internet”, insulates platforms from criminal liability, allowing them to monitor and moderate their content to create safer online spaces without taking on legal responsibility for the content they publish. In other words, Section 230 finds a compromise between the telephone and the newspaper.
Unfortunately, Section 230 has enemies on both left and right in the US, with both candidates for president in 2020 arguing it needed to be reformed or eliminated. The last word on Section 230 and other regulations affecting platforms will likely be delivered by the US Supreme Court, which has already proven itself willing to overturn significant precedent to achieve what appear to be political goals.
One solution to this mess might come from the European Union, which is establishing a track record of passing legislation that puts obligations on these largely US-based online providers. The EU’s Digital Services Act and Digital Markets Act include provisions that would increase the transparency of how social media giants amplify content. It is possible to imagine European legislation protecting the rights of platforms to remove damaging content, like neo-Nazi hate speech.
But perhaps the problem with content moderation goes beyond choosing the right metaphors to something much more basic. When the creators of early social networks and their antecedents established their terms of service, we (and I do mean we—I wrote the terms of service for one of the web’s first social media companies) decided to treat users as customers rather than citizens. We concluded that platforms could create their own rules of the road, regardless of whether users liked them or not.
The way to understand a platform like Facebook is not as one providing a service to two billion customers, but one hosting millions of conversations, each of which might benefit from its own rules and governance. A conversation between high school friends will need different rules of the road than a political debate between neighbours about a tense election. Unfortunately, in the paradigm where we create one set of rules for a space, those rules are rarely appropriate for all situations. Further, we undermine the agency that users have as citizens capable of making decisions about how online spaces should be governed.
It’s possible there is another way. Consider Archive of Our Own, a popular website created by members of the fanfic community—a form of writing in which creators expand and extend existing works, say Harry Potter or EastEnders. It was founded after its creators—predominantly women—got tired of fighting battles with the owners of existing platforms about whether their content violated copyright. (It’s generally understood that fan fiction is protected under fair use.) Not only is their site designed explicitly for fan fiction, it is governed by the community that created it, allowing for nuanced conversations about controversial issues.
It would require an enormous sea change to shift online discourse away from big social media platforms and towards a plethora of smaller ones where users can govern their own conversations as citizens, not just as customers. But with threats to online speech coming, it may be time to seriously reconsider whether we are comfortable having online conversations around the world controlled by an increasingly conservative US Supreme Court.