I recently wrote about how the Ohio State University is preventing white supremacist Richard Spencer from speaking on campus. Despite falling under the First Amendment, Ohio State is looking for a legal loophole to block speech it doesn’t agree with.

The semi-comforting news is Ohio State is subject to legal challenge. Not so in cyberspace, which Justice Anthony Kennedy called the “vast democratic forums of the Internet in general, and social media in particular.” There, access to diverse ideas is controlled by corporations and their Terms of Service. Hiding behind the bushes of private ownership, the quasi-public forums on Twitter, Facebook, Google, and their predecessors and successors, skirt the First Amendment to control what people say, read, and by extension, think. They are the censors the Founding Fathers feared.

It is time to expand the First Amendment to quasi-public institutions.

The scope of the First Amendment has regularly expanded. In the earliest days of the Republic, it applied only to the federal, and not any state, government. It wasn’t until the post-Civil War incorporation doctrine, followed by court cases well into the 20th century, that those restraints applied equally to the states. In its own founding days, Ohio State could have easily banned a speaker.

Actually, Ohio State might have been able to ban a speaker it found offensive until even more recently. It wasn’t until a 1995 case that the Supreme Court held a university’s choices on funding student publications fell under the First Amendment’s obligation not to discriminate against particular viewpoints. Other expansions of the First Amendment took place in the 1950s, when the Court extended protection to non-traditional speech, including nudity and advertising.

The problem today is when you use various web sites, you agree to a dense set of conditions, their respective Terms of Service, along with the understanding that Twitter (we’ll use it as shorthand for the range of sites and apps) can interpret things as it wishes. So while the Supreme Court continues to hold the line against banning “hate speech,” Twitter is free to apply any standard it wishes, along any political or ideological lines it wishes.

Twitter may also ban speech acting as a de facto arm of the government, skirting the First Amendment because it can. That appears to have been what happened with Twitter’s decision to ban advertising from Russian media outlets RT and Sputnik. A hazy intelligence community assessment accused both of influencing the 2016 presidential election. While the feds are stuck with that creaky old First Amendment preventing them from chasing RT and Sputnik back to Moscow, Twitter can—literally with a Tweet—silence what once were inalienable rights. Twitter followed its Russian suppression with a decision to ban whatever it thinks are “violent groups and hateful imagery and hate symbols.” Google also blocks users from their own documents saved in Google Drive if the service feels the documents are “abusive.”

Twitter has also suspended the account of conservative Roger Stone. No explanation was given, though the suspension appears to be related to Stone’s angry, expletive-filled Tweets directed at CNN. There are regular calls for President Donald Trump to be banned from Twitter for inciting violence and hate speech. A pro-left satirist found himself banned for inciting faux violence.

My own Twitter suspension occurred while defending myself against several antifa people who threatened to “punch nazis.” My response was deemed by Twitter incitement to violence (though it would not likely meet the definition the Supreme Court established—I wrote something along the lines of, “if you punch me to silence speech expect to be punched back”), and I got the boot for a week. While suspended, Twitter first forced me to delete the offending Tweet. It then allowed me read-only access, so I could see attacks continue until the antifa people tired of it all. I could not block them or respond. It felt a lot like some big guys holding me down while a bully whacked away.

Google introduced censorship in the most well-intentioned way possible: to stop child predators. The Internet giant tweaked its English-language search results to block sites it believed linked to child porn. “We will soon roll out these changes in more than 150 languages, so the impact will be truly global,” the company claimed.

While no one can argue against stopping child predators, those same tools can be used in other ways, known as the search engine manipulation effect. Google can skew search results any way it wishes to. For example, the higher an item appears on a list of search results, the more users will click on it. In a test, placing links for one candidate above another in a rigged search increased the number of undecided voters who chose that candidate by 12 percent.

Burying a link can have a similar effect. A current Google search for “greatest president of the 20th century,” for example, highlights a brisk historical debate over Ronald Reagan versus Franklin Roosevelt, and brings up over 300,000 sites. What if it yielded only one? Americans would never accept the government issuing a list of approved books for the dead tree libraries few even use anymore, but blithely accept the same from the most-used research tool in human history.

Technology has also changed the nature of censorship so that free speech is as much about finding an audience as it is about finding a place to speak. Censorship in the 21st century thus targets both speakers (example: Twitter blocks someone) and listeners (Google hides that person’s articles). There will soon be no fear that anyone will lock up dissident thinkers in some old-timey prison to silence them; impose a new Terms of Service and they are effectively dead.

The argument that Twitter and Google are private companies, that no one forces you to use their services, and in fact you are free to switch to MySpace, is an out-of-date attempt to justify end runs around the First Amendment. Platforms like Twitter are the public squares of the 21st century (seven of 10 American adults use a social media site), and should be governed by the same principles, or the First Amendment will become in practical terms largely irrelevant.

Pretending a corporation with the global reach to influence elections is just another company that sells stuff is to pretend the role of unfettered debate in a free society is outdated. Absent a court decision that places quasi-public forums under the First Amendment, we face a future of debate and discussion splintered into a myriad of ideologically-based platforms such that no one will be listening to anyone they do not already agree with. It will be a future where entities like Twitter and Ohio State protect students from the words of Richard Spencer at the expense of teaching them how to challenge those words.

Peter Van Buren, a 24 year State Department veteran, is the author of  We Meant Well: How I Helped Lose the Battle for the Hearts and Minds of the Iraqi People and Hooper’s War: A Novel of WWII Japan. @WeMeantWell