The Costs of Forcing an Online Haven for Racists Off the InternetIn late October, a gunman apparently fueled by virulent anti-Semitism and conspiracy theories fed by the president opened fire on a synagogue in Pittsburgh, killing 11 people and injuring several others. Shortly before the shooting, he had posted on Gab, a social media site that caters to “conservative, libertarian, nationalists and populist” users. Afterwards, private companies — including payment processors PayPal and Stripe, hosting service Joyent, and domain registrar GoDaddy — refused to do business with Gab, which resulted in the site being taken down for a week.In the case of white nationalists and their ilk, this unpopularity is well-deserved, but we should be very careful before we accept a world in which a few big companies can drive speakers off the internet at their discretion.As a legal matter, private companies generally have the First Amendment right to choose whom to associate with (subject to important limitations, including civil rights laws and common carrier rules). And that typically includes the right not to associate with individuals or organizations because of their virulently racist beliefs. Indeed, that’s part of how free speech is supposed to work: We have the freedom to express ideas, but others have the freedom to judge our ideas and choose not to associate with us if those ideas are noxious.Nonetheless, the more “infrastructure-like” a company is for the online world — essentially a gatekeeper to the modern-day public square — the more important it becomes that the company not drop or suspend customers on the basis of their First Amendment-protected speech. After all, nobody is proposing that people who post racist speech have their telephone service disconnected or their internet service turned off. Even those who commit crimes do not lose access to such utilities as a direct punishment.It can be hard to define who these gatekeepers are, but generally, the more vital a company’s services are to the ability to speak online and the more monopolistic it is, the greater the concern.Internet service providers like AT&T and Verizon are the clearest case: They offer a network connection, which is vital to speaking online, and exist in an uncompetitive market. They shouldn’t have the right to stop doing business with customers because of what they say over the company’s network. (That’s partly what net neutrality is all about.) Beyond ISPs, however, lies a complex ecosystem of quasi-gatekeeper companies that make online speech possible. For example, to publish a website you need not only a network connection (provided by your ISP), but also a web address (which requires a domain registrar), a place to put content (a hosting service), and protection against hacking and denial of service attacks (a security service). Advocacy groups also need the ability to raise money to support their expressive activities, which requires the use of credit card or other payment systems if that money is to be raised online.Complicating the question of whether a particular company is a gatekeeper is the fact that technologically sophisticated users may be able to perform some of these functions themselves and the fact that the company might offer many of these kinds of services in overlapping packages.Given the importance of preserving a world where anybody can post constitutionally protected speech online, the thing to watch is the degree to which there is enough competition, variety, and consumer choice among the services people need to publish online. And for many necessary services, the market is not very competitive.It may be tempting to say, “I don’t care if these racists can’t get on the internet.” But we don’t want a world in which a small number of gatekeeper companies get to decide who can and cannot speak online. The internet is the meaningful speech arena of today’s world, and anything that individuals may say in a park, they should be able to say somewhere online. Let’s remember that at various times and places in the 20th century it was generally agreed that certain speech was regarded as “beyond the pale,” including speech advocating communism, homosexuality, racial equality, and birth control. Would the GoDaddys and Paypals of the era, if they had existed, have refused to do business with individuals and organizations taking those positions?Today, private companies frequently make bad calls: Various iterations of Facebook’s and YouTube’s community standards have led to the silencing of activists of color speaking out against racism, human rights activists seeking to shine a light on ongoing abuses, and health organizations offering information about sexual health.Companies that have only the bluntest tools at their disposal are even more likely to make mistakes or reach too far. For domain registrars, for example, the only choice is to leave a website up in its entirety or take the whole thing down.Indeed, faced with controversies, companies often act in response to business and political pressures against customers who are deeply unpopular. When the major credit cards, PayPal, and Western Union blockaded Wikileaks in 2010 after that site posted government documents on the U.S. war in Iraq and other subjects, they did so under pressure from U.S. politicians, for example.The fact is, such steps may only serve to worsen the problem as racists bolster their reprehensible views with a sense of persecution. Such takedowns allow them to think, “If more Americans don’t agree with us, it’s not because our views have been rejected in a fair marketplace of ideas, but because they’ve been suppressed.”The ACLU and other free speech organizations worry about gatekeeper companies blocking even the most heinous speakers because it sets a terrible precedent without solving the underlying societal problem. If free expression is to be more than a formality in today’s world, gatekeepers to online speech should resist calls to censor or drop their users on the basis of First Amendment-protected speech.