In the wake of the Capitol riot, Facebook, Twitter and other digital platforms suspended former President Donald Trump’s accounts. Some conservatives and many free-speech advocates howled that this was a violation of the First Amendment at best, or a coordinated Big Tech attempt to suppress dissenting speech at worst. A handful of world leaders also complained, including German Chancellor Angela Merkel.

At face value, these are not unreasonable criticisms. In the 2019 case Knight First Amendment Institute v. Trump the Second Circuit Court of Appeals unanimously upheld a lower court’s decision that found it was Trump who violated the First Amendment when he blocked Twitter users who criticized him. The court’s reasoning was that his account operates “to conduct official business and to interact with the public.” Wasn’t Twitter, then, equally guilty of damaging free speech by suspending Trump? 

As critics of “cancel culture” and similar attempts to stifle dissent and debate, as well as experts on liberal democracy and electoral integrity, we offer a simple, if surprising, answer: No. 

First, no serious person really thinks free speech should be absolute and without consequences. For example, individuals and businesses can be sued for defamation or false advertising, protesters can be restricted from blaring their messages through loudspeakers at 3 a.m. and — most relevant to the case at hand — people have no right to speech that provokes a person or group to engage in violence. Landmark Supreme Court decisions have upheld the latter notion.

Second, digital platforms did not threaten free speech by flagging Trump’s untrue posts about election fraud or later banning him due to his glorification of violence. Twitter makes clear in its decision to suspend his account permanently it did so out of an abundance of caution driven by his subsequent violations of its rules and the potential for further incitements to violence in the context of the Capitol riot. The company’s reasoning had nothing to do with ordinary political speech, Trump’s campaign promises or even his lies. Nor does Twitter taking down the nearly 70,000 QAnon accounts from some of his most ardent supporters threaten free speech, either. Some of them were using digital platforms to conspire further insurrection, a crime under U.S. law.

Furthermore, if we truly care about free speech, calls to overturn or reform how U.S. law currently regulates social media would do more harm than good. Counterintuitively, social-media companies responded to Trump by following exactly what the congressional authors intended from Section 230, the law that gave birth to today’s internet. Digital platforms are empowered by this law to engage in aggressive, albeit selective, moderation. From taking down child pornography to censoring hate speech, the application of this law now rightly includes deplatforming the person who was just recently the most powerful person in the world.


Indeed, selective screening and blocking of content and users is what fosters the digital marketplace of ideas. In 1996, Congress’ Communications Decency Act — and specifically Section 230 — gave tech platforms an exemption from civil lawsuits, granting them immunity against defamation, libel and negligence. Section 230 is intended, among other goals, to promote free speech precisely by allowing these companies to moderate the content posted by third-party users, including but not restricted to, indecent content and potential criminal acts. 

The law removes the fear of civil liability that digital platforms would experience without its protections. If tech platforms are not legally responsible for what their users write and say online, then they can and should exercise discretion when removing misinformation, policing platform manipulation and curbing cyberbullying. Section 2(a) of the Communications Decency Act clearly endorses the legitimacy of “any action voluntarily taken in good faith to restrict access to or availability of material that the provider or user considers to be obscene, lewd, lascivious, filthy, excessively violent, harassing, or otherwise objectionable, whether or not such material is constitutionally protected …”

Conversely, because tech platforms are not considered publishers or even distributors, they do not have to remove posts that disgruntled parties claim are libelous, defamatory or negligent. Instead, it is the third parties who author the posts that are liable. In short, Section 230 shields platforms from having to impose blanket restrictions and engage in indiscriminate censorship, while still allowing them to curate their sites as they see fit.

Even though they are private companies without the legal obligation to free speech that applies to the government, tech platforms are encouraged by Section 230 and ensuing judicial interpretations to moderate their content to better foster the exchange of ideas. In turn, this potentially allows a more vibrant political ecosystem to flourish. Consider that users’ social-media activity has provided (true) information about candidates, promoted voter education and offered corrections to misinformation about election integrity. And even though controversial, social-media platforms allow politicians to identify and target voters; more surgical pitches increase electoral turnout and political engagement. 

At the same time, under the cover of Section 230, some unscrupulous politicians have serially deceived citizens, including the brazen lie that the 2020 election was stolen. Because digital platforms cannot and do not want to screen and curate everything, “fake news” can and does proliferate. The algorithms used by digital platforms can accentuate and more effectively spread lies and conspiracy theories, even if inadvertently, especially due to algorithmic amplification: AI making choices about what content to show users based on followers, shares and overall engagement. Translation: polarizing, sexualized and extreme videos that glorify violence and espouse conspiracy theories may draw the most eyeballs and clicks.

Yet there has always been a fraught relationship between free speech, media and politics. Deception and demagoguery are as old as politics, or at least the written word, the first vehicle for widely spreading these ills. The list of politicians who incited violence through modern means includes: Mussolini (radio/film), Hitler (radio/film), Perón (radio/television), Milosovic (television) and Rwanda’s Hutu Power (radio). And, as we witnessed on Jan. 6, speeches delivered by politicians in the flesh can be equally or even more effective than messages scribbled on the internet. 


What is not in doubt, however: Trump’s words and the subsequent actions taken by social media fall outside the bounds of free speech. Trump was at a rally outside the White House, which was filmed, and he probably committed a crime: incitement to insurrection. And then the insurrection actually happened. While arguably criminal incitement doesn’t require that any third party act on the mere suggestion, the fact that his supporters did makes this instance a much more compelling and probably easier to prove incident of incitement. This had not been true for previous episodes in which Trump used inflammatory language to stir up a crowd into a frenzy, whether on social media or in person. 

Twitter’s response was fair, proportional and prudent: Trump was initially warned and temporarily locked. He then again violated the company’s policy about glorifying violence. The nation subsequently learned about Trump’s unwillingness to quell the riot once it was in progress, which led to his second impeachment.

But Section 230 also fosters the objectivity needed to counteract the scourges of misinformation and hate. It is precisely the tech platforms’ moderation practices that eventually allow facts to surface and spread. If something like Section 230 and the commercial internet existed in earlier times, it might have been easier to arrest the propaganda advanced by leaders bent on sowing bloodshed. Regime dissenters who tried to use traditional media to counteract such vituperations were unable to circumvent the state’s censorship and repression. Fortunately today, in the U.S. case, the law has created a vibrant, albeit imperfect, marketplace of ideas with genuinely diverse viewpoints. It has also cultivated a fact checking industry that continues to improve. 

Digital platforms are developing the necessary antibodies to combat hate speech and calls for violence in a way that promotes free speech, including banning politicians who are the real threats to the First Amendment and the Constitution. The decisions by Big Tech to kick Trump off vindicates the value of Section 230. 

Of course, one might argue that social media banned Trump simply because Congress is flexing its muscle about reforming or even rescinding Section 230, or even because Big Tech is pandering to its liberal employees. There might be some truth to that view, as Twitter and other social media platforms have taken a hit by losing a big chunk of their user base and overall engagement after the former president’s deplatforming.

Even so, gutting Section 230 to remove digital platforms’ protections from civil liability for the content posted by third parties would make them much more risk averse and thus truly censorious. Before Section 230 was the law of the land, digital platforms such as CompuServe did not do much moderating at all or, more typically, they did too much of it, truly stifling viewpoint diversity and engaging in pearl-clutching prudery. If this occurred across today’s internet, facts, logic and evidence will suffer in its wake. More to the point, Twitter, Facebook and even YouTube would not exist in their current form because their business model is based on collecting, processing and selling the data created and shared through vibrant third-party engagement.


What about the view that Section 230 is a shield used by digital platforms that foment outrage through their algorithms? While there is no doubt that some social-media companies have been foot dragging on content moderation for many years and that they have the strength and deep pockets to better police their platforms, the same was once true for prior incarnations of cutting edge media distribution channels such as cable television providers that adopted voluntary systems. Getting this right sometimes takes time, learning and perhaps better regulation. Also, newspapers, magazines and even television networks can enjoy Section 230 protections if their websites allow for users’ comments.

The government could always compel social-media companies such as Twitter to stop using algorithmic amplification when offering content suggestions and also reallocate their budgets toward moderation that is more qualitative and human-based as a condition for Section 230 protections. But there are trade-offs here. Social networks can also recommend content that is edifying, and AI can help facts spread just as much as lies. And there is no one-size-fits-all way to moderate content. It can potentially involve crowdsourcing (think, mechanical Turk), one person deputized to do so, a group of anointed ethicists or improved AI. AI itself relies on human coding and intuition (think, training data sets), which suggests all moderation requires planning, judgment and learning.

A better use of Congress’ time if they are worried about technology and democracy would be to promote civic education and provide broadband to all Americans to help both spread accurate information about how elections are conducted and debunk conspiracy theories by helping to disseminate facts on, yes, digital platforms. The truth is, the demand-side factors driving misinformation and conspiracy theories will endure if the commercial internet as currently constituted disappeared tomorrow. The answer to the rampant fear, distrust, polarization and uncertainty about a fast-changing world is not to ban the messenger but to do something about the message. Policymakers would be wise to focus on those who have been left behind by globalization, racial injustice and ignorance.