The technology sector has been a bright spot during the coronavirus pandemic. Online platforms have allowed millions of Americans to work from home, while rising share prices of major tech firms have fueled much of the stock market’s rebound.

If the tech industry thought any of this would buy them political goodwill, however, they were mistaken.

On Thursday, President Trump signed an executive order to curtail protections for the industry under a provision of federal law known as Section 230. Even as Republicans have been pushing to shield businesses from legal liability stemming from COVID-19, Trump’s order removes liability protections from social media platforms.

His action is certain to face legal challenges. Section 230 is an essential underpinning of free speech law on the Internet. In fact, it’s often been described as “the 26 words that created the Internet.”

“Amending Section 230 is always a challenging question because it has been such a bedrock in how online speech has developed in this country,” says Emma Llansó, director of the Free Expression Project at the Center for Democracy and Technology, a civil liberties group.

Trump’s executive order has three main components. The Commerce Department will ask the Federal Communications Commission to create new regulations to clarify when an Internet platform had violated the good faith provisions of Section 230. The order directs the Federal Trade Commission to review White House complaints about political bias and potentially sue companies for violations. And it instructs the federal Justice Department to consult with state attorneys general about allegations of political bias.

“Big Tech is doing everything in their very considerable power to CENSOR in advance of the 2020 Election,” Trump tweeted on Wednesday. “If that happens, we no longer have our freedom. I will never let it happen! They tried hard in 2016, and lost. Now they are going absolutely CRAZY. Stay Tuned!!!

Twitter has long been one of the president’s primary megaphones, allowing him to bypass traditional media outlets (which obsessively cover his tweets) and reach more than 80 million followers directly. Trump’s executive order appears to have been precipitated by Twitter’s decision on Tuesday to append fact checks to a pair of his tweets that made dubious claims about mail-in voting. On Friday, Twitter flagged a tweet posted by Trump for violating its rules regarding "glorification of violence." Trump referred to protesters in Minneapolis, where there have been looting, arson and vandalism this week, as "THUGS... Any difficulty and we will assume control but, when the looting starts, the shooting starts."

While it may be ironic that the president is going after one of his own major platforms, some of the ideas in his order have been percolating for months. “No longer are tech companies the underdog upstarts; they have become titans of U.S. industry,” U.S. Attorney General William Barr said in a widely noted speech in February. “Given this changing technological landscape, valid questions have been raised on whether Section 230’s broad immunity is still necessary, at least in its current form.”

Even as social media companies find themselves under attack by the administration and Republican senators such as Marco Rubio of Florida and Josh Hawley of Missouri, they have their critics on the left as well. Democrat David Cicilline of Rhode Island, who chairs the U.S. House Antitrust Subcommittee, has been working on legislation to limit Section 230 immunity for platforms that knowingly publish false political ads.

“On the left and the right, people are saying these companies are discriminating against us,” says Grace Morgan, director of external relations at the Taxpayers Protection Alliance, a government watchdog group that defends Section 230. Like others quoted in this story, she was interviewed about the law prior to Trump’s order being floated.

Social media platforms find themselves criticized either way, whether they’re policing speech or failing to do so. Trump’s order may not survive in court, but the so-called techlash will continue.

What Is Section 230?

Section 230 takes its name from the 1996 Communications Decency Act. Under that law, telecommunications providers could face charges if they fail to take “good faith, reasonable, effective and appropriate actions” to prevent minors from seeing pornography or indecent content.

Section 230 offered platforms that do make a good faith effort protection from liability. It also makes it clear that “no provider of an interactive computer service shall be treated as the publisher and speaker of any information provided by another information content provider.” In other words, Internet platforms are not treated like traditional publishers. It would be one thing to expect a magazine publisher to vet its own articles, but another thing entirely to expect social media companies to check millions of posts per day.

When the law was enacted, Congress was thinking less about the infant Internet than traditional broadcasters and telecoms. But Section 230 allowed the Internet to flourish — not just social media giants such as Twitter and Facebook, but any site or Wiki or message board with a comment section. Plus any site that features reviews. Really, just about everything we think of as the Internet.

“Taking away 230 protections would limit the interactions you’d be allowed to have with people online,” Morgan says. “When people think about bias, they don’t think about how taking away 230 would stop people from reviewing a product on Etsy or rating a property on Airbnb or commenting on mom blogs.”

How 230 Became a Target

In 2018, Trump signed the first major revision to Section 230. Known as FOSTA-SESTA, the law targeted online sex trafficking and prostitution, making platforms responsible for content promoting such illegal activities.

Some state attorneys general viewed FOSTA-SESTA as just a start. The National Association of Attorneys General has repeatedly sent letters to Congress complaining that courts have interpreted Section 230 in ways that protect individuals and companies that benefit from illegal online activity. “While helpful, (FOSTA-SESTA) does not allow for state prosecution of other crimes involving these platforms, such as online black market opioid sales, identity theft and election meddling,” Kansas Attorney General Derek Schmidt said last year.

Attacking Section 230 has been part of a broader campaign state AGs have launched against tech, says Paul Nolette, a political scientist at Marquette University. A group of attorneys general, led by Ken Paxton of Texas, has been investigating Google in preparation for a likely antitrust lawsuit.

“On the Section 230 issue, it is consistent with the increasingly aggressive approach AGs have been taking against the tech sector,” Nolette says.

Section 230 has become a convenient shorthand for politicians wanting to attack the tech industry, suggested Adam Conner, vice president of technology policy at the Center for American Progress, a progressive think tank. People may have lots of issues with the big tech companies, but content they find objectionable is a harm they can feel viscerally. “Section 230 is not well understood, but it comes a convenient shorthand for a lot of problems, not all of which are tied to it,” Conner says.

Putting Limits on Protection

In March, a bipartisan group of senators introduced the EARN IT Act, which would require Internet companies to actively fight child exploitation or risk losing some liability protections under Section 230.

There are problems with such an approach, according to Internet speech experts. Even laws meant to address narrow issues, such as sexual abuse of children or pornography, can present technological challenges. It’s relatively easy to filter such material — but not actually easy. Broader limitations on Section 230, including the questions of bias raised by Trump’s executive order, present much greater challenges.

Let’s say calling a politician “dumb” was considered evidence of bias. Someone could post comments that don’t contain words such as “dumb,” “idiot” or “stupid.” Instead, they could suggest the idea obliquely, for instance by offering a twist on a phrase such as “a few pounds short of a full load,” coming up with variations that would make the point clearly to human readers but never be flagged by machines.

What about intent? At midnight on Thursday, Trump retweeted a video featuring Couy Griffin, a county commissioner in New Mexico, stating that “the only good Democrat is a dead Democrat.” Griffin said he didn't mean the words literally, but he has also said Democratic governors should be executed for treason. Who gets to decide when an apparent threat is being made sarcastically, and whether that matters?

Regardless, the fact is that major Internet platforms such as Facebook and Twitter are far likelier to have the enormous resources needed to build artificial intelligence systems and hire armies of humans to root out offensive or harmful speech than their smaller competitors.

“The irony is that in weakening Section 230, it’s not necessarily Facebook and Google you’re going to hurt,” says Elliot Harmon, activism director at the Electronic Frontier Foundation, a civil liberties law firm. “You’re going to hurt the ability of new companies to compete.”

The Perils of Regulating Speech

Free speech protections under the First Amendment are meant to protect individuals from restrictions placed by the government. Private companies don’t have to allow free speech. That’s why employers can block their workers from showing up at political rallies, or stores can post “no soliciting” signs on their doors.

It’s the protection from liability that encourages Internet companies to monitor illegal activity. If something is a crime, it’s still a crime on the Internet. Section 230 lets companies moderate or remove content, assuming they’re making a good faith effort, even if they’re not catching everything, Conner says.

Otherwise, they might issue a blanket policy that they bear no responsibility for content and fail to block any infractions, allowing more indecent or offensive material to flourish. Conversely, it could prompt platforms to curtail speech of any or all users, in order to avoid legal peril.

Facebook CEO Mark Zuckerberg made clear his preference for a hands-off approach in an interview Wednesday with Fox News. “I just believe strongly that Facebook shouldn’t be the arbiter of truth of everything that people say online,” he said. “Private companies probably shouldn’t be, especially these platform companies, shouldn’t be in the position of doing that.”

Of course, that stance is part of what makes Facebook unpopular in some quarters. The site has clearly been used to spread misinformation, not least when it comes to election interference.

So, should Internet platforms have no standards at all? If they should, who should set those standards? Free speech adherents say it would be a mistake for government agencies to determine which sites are prone to bias.

“The order directs federal agencies to ignore judicial precedent and cut online ad spending with certain social media platforms,” Daniel Castro, vice president of the Information Technology and Innovation Foundation, said in a statement on Thursday. “This is an inappropriate use of the powers of the presidency and opens the door to legitimize similar actions by other political leaders around the world. The long-term impact of such retaliatory policies would be a chilling of free speech online, especially counter-speech, a rise in disinformation, and a decline in innovative social media platforms.”

Trump’s order is rooted in the sense expressed by many conservatives that they’re treated unfairly by social media platforms. In 2018, Facebook apologized to the conservative group PragerU for blocking several posts and videos under the initial misapprehension that they violated its community standards policy.

Morgan says accusations of anti-conservative bias are mistaken. She says that many conservative voices have gained prominence thanks to social media, while her own group’s work seeking to expose wasteful spending and regulations wouldn’t be disseminated as widely without it.

“Government regulation, from our perspective, is never the answer,” Morgan says. “If we took away 230 protections, it would allow social media companies to take down anything they please.”