Twenty years ago, a federal appeals court said Kenneth Zeran couldn’t sue AOL for failing to remove defamatory posts. It is no exaggeration to say that had the court gone the other way, much of today’s Internet could not exist in its modern form. But when the issue is copyright instead of defamation, Congress and the courts have resisted this lesson; instead of nurturing new industries, they’ve snuffed them out. And just as it was impossible to guess in 1997 the many platforms, tools, and communities that would emerge after the Zeran v. AOL decision, it is impossible to know now how many innovative industries will never emerge due to its copyright counterparts.

Zeran answered a critical question for online services: If a user posts something that’s defamatory, and so violates the law, is the service liable? It’s easy to see why the answer must be no. An online community like AOL doesn’t work without content contributed by users; without that content there is no community. That was true in 1997, and it is even truer today, when content generated by users underlies all kinds of online services. But if a company had to police every piece of user-generated content or, worse, were liable every time a user went too far, it would be impossible to run online services at scale. Facebook and YouTube couldn’t vet each post for defamation liability; certainly a 10-person startup couldn’t do so.

So Zeran made it possible for online services to exist without incurring crippling liability. And that led to a surge of services that billions of users rely on, an explosion the scale and breadth of which the court could not have imagined in 1997. Some of these services are straightforward descendants of AOL and its contemporaries: discussion forums, search engines, and blogging platforms all resemble tools that existed in 1997. Others were less foreseeable. Social networks and online video existed in 1997, but the sheer variety and scope of platforms like Facebook, YouTube, and Snapchat would surprise someone from 1997; likewise, tools like Wikipedia, Genius, and Adblock Plus, which rely on content contributed by users, had few parallels when Zeran was decided.

Zeran’s rule of limited liability was a public-policy success because it created free space in which whole industries could develop. The same cannot be said of copyright policy on the Internet; if Zeran had sued AOL for failing to take down copyrighted content instead of defamatory content, he probably would have won. So new business models that involve copyrighted content are at much greater risk than those involving other kinds of user-generated content.

Take the case of Aereo, a service that let users watch broadcast TV on the Internet. Courts had long held that consumers can legally copy works for space-shifting and time-shifting, so they can watch and listen to video and music at different times on different devices. This is what iPods and DVRs do, and it’s also what Aereo promised to let users do. Aereo set up individual, dime-sized antennas for users so they could record and stream broadcast TV channels. This system was just a remote DVR: instead of recording shows onto a hard drive in her home, a user could outsource that function to Aereo, just as she might outsource email or file storage to an online service. And so the U.S. Court of Appeals for the Second Circuit held that what Aereo did was legal, just like any other DVR would be.

The Supreme Court disagreed, in a funhouse mirror image of Zeran that destroyed innovation instead of encouraging it. The court noted that Aereo marketed itself as a replacement for cable TV and reasoned that since cable companies “perform the copyrighted work[s] publicly” when they transmit them to subscribers, Aereo must do so as well. It didn’t matter that Aereo’s transmissions were triggered by users, not Aereo itself, or that each user recorded and transmitted her own copy from her own antenna, or that the transmissions were available only to the user, not to the broader public. Instead of analyzing these key distinguishing features of the Aereo system, the court adopted what Justice Antonin Scalia, in dissent, called a “looks-like-cable-TV” test: if a company creates a new business model that competes with an incumbent technology, courts should bend the law to apply the same copyright rules to each. So while in Zeran the court took a narrow view of the plaintiff’s rights, requiring him to sue the people who posted defamatory content instead of the platform hosting that content, in Aereo it took the broadest possible view of the plaintiff’s rights.

Did the Aereo decision actually prevent any innovation? It’s impossible to tell for sure, but there are all sorts of possible business models that would run afoul of the court’s rule. One big contender would be a service to solve fragmentation in video streaming. When all TV was broadcast over the air, people could buy any TV set and pick up any show on any channel. As video moves online, though, there is no streaming service that has every show and no box that can run every streaming service. Instead of just changing the channel, today a user might have to skip a show if she doesn’t have a box that can play it. It’s easy to imagine a service, then, that could tune in and stream video from any service to a custom app or a web browser—effectively, space-shifting for streaming services. But under Aereo, that service is probably illegal. The result is that incumbent rights-holders can veto new businesses that might threaten their incumbency, a power they have been happy to exercise.

Copyright holders have long used their copyright monopoly—legally—to prevent competition, but they have been constrained by limitations like the first-sale doctrine. Back when Blockbuster Video was the state of the art in watching movies, studios couldn’t stop stores from renting them, since the law blocks a copyright holder from restricting what someone does with a copy after it has been sold. But the shift to online business models has upset this balance between creators and others using those creations, since online streaming inherently creates copies and so isn’t subject to the first-sale doctrine.

The Aereo court could have helped restore the balance between creation and competition by limiting rights-holders’ powers, letting people use online services to do the same things they have long been able to do offline. This would have encouraged entrepreneurs to create valuable new businesses and services, just as the Zeran decision did two decades earlier. Instead, the court went the other way. The fault may lie more with Congress than with the courts, since in Zeran, Congress had created an express immunity for businesses relying on user-generated content; Congress’s similar immunity for copyrighted content, a safe-harbor rule that applies when sites have notice and take down allegedly infringing content, is much more limited. Still, there is a long history in copyright law of technologies that look like pirates at first but eventually become respected businesses; recorded music, the VCR, even sheet music were all at one point seen as threats to rights holders. Congress and the courts should keep the Zeran lesson in mind before backing away from that history and preemptively killing off the online services of tomorrow.

Roger Allan Ford is Associate Professor of Law at the University of New Hampshire School of Law and Faculty Fellow at the Franklin Pierce Center for Intellectual Property. Thanks to Alexandra Roberts, Ann Bartow, Eric Goldman, Jeff Kosseff, Jennifer Berk, John Greabe, Keith Winstein, Leah Plunkett, and Ryan Vacca for comments on a previous draft.

This essay is part of a larger collection about the impact of Zeran v. AOL curated by Eric Goldman and Jeff Kosseff.