For Airbnb, suing American cities had become something of a monthly ritual. It started in late June 2016 with a federal lawsuit against San Francisco, the company’s hometown, over an ordinance meant to crack down on illegal short-term apartment rentals. About a month later, Airbnb sued the city of Anaheim for much the same reason. In September, it followed up with a salvo against Santa Monica, and finally in October, Airbnb sued the city and state of New York. The specifics differed in each of the cases, but they all hinged on a common variable: Lawmakers had passed housing regulations that could hold Airbnb accountable for illegal listings on its website. The home-sharing giant was now in a position where it had to either ensure that its users were following the law or face hefty fines. Granted, Airbnb is no stranger to hostility from housing authorities, and it has often been willing to sit at the table with local governments to hash out differences. But this was one concession the company was not willing to make. Pay fines for errant postings? Never.

After all, it has the bulwark of cyberlaw on its side. In each of the four lawsuits, Airbnb’s lawyers confidently buttressed their defense with a 20-year-old federal statute: Section 230 of the Communications Decency Act. Tucked into the mammoth Telecommunications Act of 1996, this landmark piece of legislation is often cited as the most important tool ever created for free speech on the internet. It includes a crucial “safe harbor” provision that gives online platforms legal immunity from most of the content posted by their users. The money quote goes like this:

No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider.

With those 26 words, the federal government established the regulatory certainty that has allowed today’s biggest internet companies to flourish. Without Section 230 — the popular theory goes — there could be no Facebook, Amazon, or Twitter. Yelp’s one-star reviews would have rendered it helpless against litigation from angry business owners, and Reddit’s anonymous trolls would have long ago invited a barrage of devastating libel lawsuits.

In a nutshell, Section 230 is the statutory glue behind everything you love and hate about the internet.

But as Airbnb’s recent court battles highlight, the scope of the law is far from settled. Stark attempts by legislators and judges to refine or redefine Section 230’s boundaries are chipping away at the broad immunity websites once took for granted, and some defenders of Section 230 worry that the core protections enshrined in the law are under threat.

The changing tide is evident elsewhere in the courts, where recent rulings have dealt blow after blow to internet companies that depend on safe harbor. In June, for instance, a judge ordered Yelp to remove reviews that were found to be defamatory. In August, Twitter was told it could not use a Section 230 defense in a lawsuit over unwanted texts. And in September, a panel of federal judges greenlit a $10 million lawsuit against Match.com brought by a woman who was stabbed by a man she’d met on the site.

“At some point you wonder if this is just judicial activism going crazy,” says Eric Goldman, a longtime cyberlaw expert and professor at the Santa Clara University School of Law. “You wonder if judges are saying, I know what Section 230 says. I just don’t agree with it.”

To understand how Airbnb’s legal fights underscore this rift, start with its own hometown of San Francisco. Earlier this year, that city’s Board of Supervisors unanimously approved an ordinance meant to force home-sharing websites like Airbnb to remove short-term listings that were not registered with the city. Failure to do so could mean fines of up to $1,000 per day. The ordinance, introduced by Supervisor David Campos, tested the limits of Section 230 immunity in an interesting way. Because apartment listings on Airbnb are posted by third parties, conventional cyberlaw wisdom dictates that Airbnb can’t be seen as the publisher of those listings.

Wisdom aside, San Francisco has an intractable housing shortage to think about, and home-sharing platforms are seen as exacerbating the problem. One city estimate found that 80 percent of local short-term rental hosts on Airbnb did not bother to register their units. So the city, led by Supervisor Campos, decided to put the responsibility on the platforms themselves. It’s easy to see why Airbnb would turn to the courts to fight this: If it has to pay a fine for every scofflaw who operates an illegal rental, its business model quickly falls apart. And the ordinance appears to be a clear violation of Section 230 — at least on its surface. Airbnb is not responsible for third-party posts. End of story.

But officials in San Francisco think they’ve found a creative workaround, one they said makes the Section 230 argument irrelevant. Carolyn Goossen, a legislative aide to Supervisor Campos, called me up one afternoon to explain. “The city is regulating a business activity — not content or posts,” she says of the ordinance. “It says a hosting platform cannot conduct business with a short-term rental unit if that unit has not registered with the city. If it does conduct business with them, then they are subject to a fine.”

In other words, Airbnb can host whatever illegal listings it wants. It just can’t make money from them as a booking service. Somehow, I suspected that this workaround would not appease Airbnb, whose revenue — estimated at $900 million in 2015 — is dependent on its ability to take a percentage of its hosts’ rental fees. I was right. Alex Kotran, a spokesperson for the company, says the amendments introduced by Campos do nothing to resolve the legal shortcomings addressed in the lawsuit. Although he couldn’t speak in detail, he gave me a list of legal experts who could weigh in on the Section 230 implications of the lawsuit. Coincidentally, one of those experts was Eric Goldman. When I mentioned Campos’s rationale to him, he scoffed.

“That sounds more like political spin than a legal analysis,” Goldman told me in an interview earlier this year. “In the end, no matter how it’s phrased, San Francisco wants to deputize Airbnb as its assistant tax collector. That fundamental effort of putting Airbnb in the role of policing what its users are doing is the kind of thing that Section 230 was designed to prevent.”

Except that James Donato, a US District Court Judge for California’s Northern District, didn’t see it that way. In November 2016, he dealt a major setback to Airbnb when he rejected the company’s request to block the ordinance. Donato didn’t buy Airbnb’s Section 230 argument. As he put it, San Francisco’s ordinance doesn’t treat Airbnb as the publisher of illegal rental listings, nor does it force Airbnb to police its website and remove such listings. It simply holds Airbnb accountable for its own conduct: providing “booking services” in connection with unregistered units.

“As the text and plain meaning of the Ordinance demonstrate, it in no way treats plaintiffs as the publishers or speakers of the rental listings provided by hosts,” Donato wrote.

For supporters of Section 230, it was the next domino to fall. In a blog post the following week, Goldman wrote that the ruling could put all online marketplaces at risk. What if city governments were to, say, require Amazon to verify that its vendors have local business licenses? And what happens when other cities look to replicate San Francisco’s model? The internet, and all that we’ve come to take for granted about it, is looking shakier all the time.

Everything you love and hate about the internet was born in a segregated lunchroom. It was in Washington, D.C., in the spring of 1995, where Congress came to eat — a well-maintained buffet area located in the south wing of the US Capitol building. Here, Republicans and Democrats shoveled food onto their plates and then dispersed to go sit with their own kind.

Two congressmen sought to break the pattern. One afternoon, Chris Cox, a Republican from California, and Ron Wyden, a Democrat from Oregon, put their plates together and strategized how they might cut through the frosty partisanship that divided the Hill. They agreed that the way to get bipartisan support for a cause was to focus on the future, on some pressing problem that lacked the baggage of the usual issues like abortion or taxes.

And in 1995, that pressing problem was the internet. The fledgling network of computer systems was regulated by a patchwork of creaky laws written for an earlier era. Legislators didn’t get the internet. Neither did judges.

So fragile was the whole thing that a single court ruling in May of that year threatened to smother the internet in its crib. Prodigy, an early provider of online services, was found to be legally liable for a defamatory anonymous posting on one of its message boards. The ruling had chilling implications: If websites could be sued over every piece of content that someone didn’t like, the internet’s growth might come to a halt. Cox read about the Prodigy ruling on a flight from California to Washington and had one thought: I can fix this!

“A light bulb went off,” he told me recently. “So I took out my yellow legal pad and sketched a statute. Then I shared it with Ron.”

That statute eventually became Section 230. In hindsight, the concept is ridiculously simple: Websites aren’t publishers. They’re intermediaries. To sue an online platform over an obscene blog post would be like suing the New York Public Library for carrying a copy of Lolita. For a young internet facing a potential avalanche of speech-squelching lawsuits, Cox and Wyden’s provision was a creative workaround — a hack — that allowed this new form of communication to grow into the thriving network of commercial enterprises we know today.

“The internet would look very, very different,” Cox says.

Given how often Section 230 is championed cited, and showered with superlatives, you might not know there is a raging debate going on about how well the law actually works. Running counter to the praise from free speech advocates is a chorus of legal scholars and civil rights activists who point to its flaws and question how it’s been interpreted by courts. Did the law really intend, for instance, to provide a safe haven for unscrupulous landlords who flout housing regulations? Do tech companies really have no responsibility for the havoc — including vicious behaviors such as online bullying, doxing, and death threats — let loose by their platforms?

One person who has been asking these questions is Mary Anne Franks, the legislative and tech policy director for the Cyber Civil Rights Initiative, a group that fights online harassment and offers support for victims whose lives have been wrecked by it. Franks is a professor at the University of Miami School of Law who speaks in long, elegant sentences that somehow manage to reveal moral clarity through precise legal language. In a recent interview, she told me she found Section 230 absolutism troubling, particularly because so many of its defenders seem to accept the negative consequences of the law as an unfortunate tradeoff for free expression.

“That kind of ranking of values is strange,” she says. “The implicit judgment that’s made when people say ‘tradeoff’ is that we can’t do better, and I just don’t believe that.”

Legal immunity, Franks says, is great if you’re a tech company that has been able to thrive under it, but not so great for those who have suffered at the hands of anonymous trolls. That raises the question of whether Section 230 really works for every internet user—or just a certain kind of internet user.

“If you belong to a certain type of demographic — and to be blunt about it, if you belong to a white, male, fairly privileged demographic — you probably have a pretty good experience online and think overall this is working out quite well,” Franks says. “But if you ask the opinions of people of color, or women, people whose lives have been turned upside down by the kinds of harassment that technology makes possible, I think it’s a different story.”

She agrees that Section 230 has done a good job of distinguishing between intermediaries and producers of content. Mark Zuckerberg can’t be considered the publisher of every Facebook status update, and we can all agree he shouldn’t be. But she challenges the popular notion that the law has fostered a more free and open internet — and it’s not necessarily because of the law itself. Rather, she says the courts too often favor immunity.

“When you have that much misapprehension about what Section 230 does, maybe we just need to clarify,” Franks says. “I’m not opposed to that, but I do wish courts would do more with what they have and not grant immunity to everyone who claims it.”

Back in Santa Clara, Goldman has a different take. He says the past year or so has seen a disquieting number of courtroom losses for Section 230. He recently cataloged more than a dozen of them on his blog, including legal blows against Facebook and Google, in addition to the aforementioned ruling that ordered Yelp to remove certain reviews. That last one really gets him going. “It’s just wrong, wrong, wrong, wrong, wrong,” Goldman says.

As an advocate for internet speech, Goldman is about as knowledgeable as they come. His blog is an exhaustive repository of Section 230 information, with commentary and links to cases dating back to 2005. He’s been teaching internet law since before Section 230 was passed and wrote a paper in law school about user-generated content before that term existed. To understand why he considers internet speech so vital, he says just look at what the world was like in the pre-internet era, when interested participants in society had no immediate outlets through which to speak up and share their voices. He talks about his first exposure to online bulletin boards in the early 1990s as if it were a spiritual awakening.

“All of a sudden there were these communities, and I could become an equal and vibrant participant in them,” Goldman recalls. “I thought that was amazing. It’s what I always wanted, and never knew it existed.”

But free speech advocates like Goldman are not the biggest reason Section 230 remains so fiercely guarded after 20 years later. Safe harbor has allowed the modern internet to flourish, which means it has also enabled the most powerful companies in modern history — with the best lawyers money can buy. If Silicon Valley is the capitalist equivalent of Superman, Section 230 is its yellow sun, the source of invincibility for Google, Facebook, Amazon, Twitter, and all the rest.

Lawyers for tech companies litigate ferociously to prevent even a sliver of erosion in the law. Lobbyists work tirelessly to block legislation that would undermine safe harbor. What began as a provision to promote the growth of an emerging technology is now a legal tool to protect the business interests of the powerful. By any measure, the internet is no longer that infant at risk of being smothered. “The argument now is, look, the internet is pretty robust,” Franks says. “The idea that a regulation here or there is going to bring down the internet isn’t that plausible. I’m not sure it was plausible in the 1990s.”

Ron Wyden left the House of Representatives not long after Section 230 passed, having won a special election to the US Senate in January 1996. He’s still serving in the Senate today. He recently spoke with me from Oregon about the original intent of Section 230 and whether he believes the courts have properly interpreted it over the years. For the most part? Yes, he says, but he never imagined it would go as far as it did. Who could have?

The Cox/Wyden amendment, as Section 230 was initially known, was a direct response to the Senate version of the Communications Decency Act, an anti-pornography bill introduced by the late J. James Exon, a Democratic senator from Nebraska. Exon was worried about how the emerging World Wide Web was making it easier for minors to access obscene material. To solve the problem, he took a censorship approach, introducing a bill that critics said was both implausible and unconstitutional. It still passed the Senate overwhelmingly — because who is going to vote against keeping kids away from smut?

Wyden and Cox put up a fight. They introduced a House amendment, which also passed overwhelmingly. The final version of the Communications Decency Act included both Exon’s input and Cox and Wyden’s input, but it didn’t stay that way. A year after it was passed, the Supreme Court struck down Exon’s portions of the law. Section 230 remained.

Asked his thoughts about the impact of Section 230, Wyden says mostly what you would expect: He truly believes, as does Cox, that the internet would look very different without it. But one thing about his answer surprised me. The reason Section 230 came to be in the first place was not just so that websites could leave objectionable material up. It was so they could take it down. The Prodigy court ruling that sparked Cox’s light bulb all those years ago hinged on Prodigy’s decision to moderate its message boards. Put another way, Prodigy enforced content guidelines and removed posts that violated them, which is pretty much what all websites do nowadays. But because Prodigy exercised some degree of editorial control over its content, the court saw it as a publisher — legally liable for whatever appeared on its site.

That ruling put online platforms in a strange position. They could either let users post whatever they wanted, or they could enforce content guidelines and risk getting sued. Wyden says one of the things that is not well understood about Section 230 is that it was really meant to address this contradiction.

“Before Section 230, online content providers left offensive material up because of the liability they would incur from pulling it down,” he tells me. “I think it fleshes out the debate for that to get mentioned, because hardly anybody mentions that.”

It also presents an interesting irony in some of the Section 230 court battles happening right now. Airbnb isn’t fighting with San Francisco for its right to take down illegal postings, nor is it fighting to leave them up. It’s fighting to have it both ways. On the one hand, it says it can’t be responsible for policing what its users do. On the other hand, it needs to exert enough control over those users to create and enforce a sophisticated system of rules and policies — one trustworthy enough that millions of people will use it to invite strangers into their homes. It’s a delicate balance that highlights how blurry the line between intermediary and publisher can be.

When Judge Donato denied Airbnb’s request to block San Francisco’s ordinance in November, he appeared to draw yet another boundary for Section 230. Now it would be up to Airbnb to push back. But a week after Donato’s ruling, something interesting happened: Airbnb gave in. The company agreed to work with San Francisco to create a mandatory registration system for its hosts. Chris Lehane, Airbnb’s global policy chief, told the San Francisco Chronicle it was a serious proposal to “once and for all address the core issues that exist in San Francisco.” The city, for its part, will not enforce the ordinance for the time being. A settlement conference for the lawsuit is scheduled for this month.

Maybe Airbnb smelled a legal loss. Or maybe it just wanted to move forward. I reached out to the company to ask if it still believes the ordinance, as written, violates Section 230. I didn’t get a response.

One of the reasons this legal fight resonated with me is because it sums up the often acrimonious struggle to balance the digital with the physical. Section 230 regulates the internet, but it would not mean a whole lot if it didn’t have implications in the real world. And what’s more physical than our homes and neighborhoods?

During my conversation with Franks, we talked about this clash, and how it boils down to a larger philosophical question about whether the law should inherently treat the internet differently. It’s true that housing regulations don’t always make sense, sublet clauses are confusing, and many building codes can seem obsolete—but they were written with local communities in mind. Strip them away and you may lose more than you bargained for.

“The reason why zoning laws are there to begin with is because calculations were made at a certain point, in certain cities, that they wanted certain rules to be in place because there was a certain quality of life,” Franks says. “If the internet is allowed to basically break all that up, we’re giving it a power that I’m not so sure it deserves.”

Creative Art Direction: Redindhi Studio
Illustrations by: Lauren Cierzan