Why Joe Biden and Donald Trump are both wrong about Section 230

In the course of the presidential race, Joe Biden and Donald Trump both advocated repealing Section 230, and politicians across the spectrum want to replace it with more detailed regulations designed to address the political influence of social media and technology platforms. I think that's the wrong approach: Section 230 is exactly the kind of evolutionary regulation that we need to govern technology and its use.

Simply put, Section 230 says that social media platforms and other online services are not legally responsible for content posted by users. This brief amendment, the heart of which is 26 words long, has protected hundreds of companies and users from being blamed for what others do online, and freed platforms to moderate content without being sued.

Now, some politicians want to change or repeal Section 230 to punish large tech companies for alleged bias against certain content or for spreading misinformation. Legislators have proposed at least 10 bills in the past 18 months to modify it. In May, alleging that social media platforms are biased against Republicans, President Trump issued an executive order that has prompted the Federal Communications Commission to consider revising Section 230 through regulations.

Yet Section 230 is a perfect example of why most technologies are better governed by applying simple, general principles on a case-by-case basis. We don't need complex, detailed legislation or regulation: Creating it will only freeze the rules based on current technology, bogging down future innovators.

Understanding how Section 230 works requires understanding two different types of law: statute law and case law. It's worth taking the time to do so here because Section 230's power comes from its case law — a legal approach most people forget exists.

An evolutionary approach

The actual text of Section 230 (a 1996 amendment to the Communications Act of 1934) is statute law, which is what most people think of when they think of "a law": a set of rules written by legislators. Analogizing to software development, statute law uses the waterfall methodology: It's centrally managed, sequential development with infrequent releases.

Compare how courts have generated volumes of Section 230 case law (sometimes called "common law"). Judges create case law by applying lessons from past cases to new situations; those new decisions then inform other judges in future cases. Core doctrines, such as negligence or nuisance, develop over time. Those doctrines are often summarized in a judge's opinion, but they are not written down in any one location. Instead, the law is the collective reasoning and conclusions of many court cases, each of which resolves a specific conflict. Case law is more like agile software development: Many decision-makers frequently release small changes and adjust as new issues arise.

For well-understood problems and slow-changing industries, statute law can work well. However, poorly designed statutes can impose unnecessary requirements that don't solve the intended problem, and outdated statutes can be difficult or counterproductive to comply with. Furthermore, statute law rarely changes because legislation moves very slowly. Usually, statute law just accumulates, unchanged, over time. Legislators' political agendas or special interests can also distort the design of statutes.

By contrast, no one designs case law: It evolves. When a court applies a principle from an earlier case, it also shapes that principle for future courts. As a result, compliance with case law is sometimes less clear-cut. Companies often estimate it as "liability risk" — how likely are we to be sued for this? And different judges might interpret and apply the same case law differently.

But this ambiguity also means that over time, case law can evolve to address new problems. Case law focuses on actual, existing problems, not theoretical or future problems. This focus means that useless case law usually fades out or is expressly eroded by court decisions over time. And special interests cannot easily distort case law because of its decentralized nature and its focus on specific conflicts rather than entire industries.

Because case law is evolutionary, it's better suited to regulating fast-moving technology than detailed legislation, which can quickly become outdated.

Rather than wait until someone is hurt and can sue, it might seem better to pass a law ahead of time to prevent potential harms. This makes sense when Congress understands the future path of a technology well and when the harms are likely to be widespread, catastrophic and irreversible. But in most cases, even experts don't understand the future of a technology very well, let alone legislators. While detailed legislation may prevent some harms, it can also prevent beneficial uses and will likely miss other harms. In these cases, it's better to address problems as they arise.

To return to the software development analogy, using the waterfall methodology to diagnose and solve every potential need or problem before releasing the software might make sense for, say, a program that will control a nuclear plant. However, agile development is often a better way to discover what people really need and what problems they actually face.

The moderator's dilemma

Section 230 is unusual because it's a statute that has some of the strengths of case law. In fact, Section 230 is itself a statutory solution to a conflict in case law about defamation.

At the time Section 230 was passed, some courts had created "the moderator's dilemma": Early online platforms such as AOL and Prodigy worried that taking even small steps to moderate content, such as removing spam or porn, would make them legally responsible for all third-party content that remained on their platforms. Many worried that such forums would either become free-for-alls with no moderation whatsoever, or highly curated spaces where only approved posts were allowed.

Congress altered the evolution of case law by enacting Section 230, solving the moderator's dilemma and unleashing a wave of innovation in online platforms. But in doing so, Congress also opened a new conflict. The courts, now bound by Section 230, continue to apply the law on a case-by-case basis. Today, Congress faces political pressure to modify the statute, and the Trump White House has added a new wrinkle by bringing the FCC into the mix.

Some argue that, in disrupting a still-evolving common law principle, Congress acted hastily in passing Section 230. However, some hasty ideas are still admirable – or fortunate – in their foresight. Section 230 captured the commonsense notion that people, not the tools they use, are responsible for what they do online. Unlike most statutes governing technology, Section 230 is short and doesn't put some administrative agency in charge of passing reams of detailed rules. Instead, Section 230 wisely charged courts with interpreting it on a case-by-case, and thus retained many of the benefits of common law. It has remained flexible and adaptable while also setting forth a clear standard.

Section 230 works, in my view, because, way back when Mark Zuckerberg was still in grade school, Congress didn't try to solve the moderator's dilemma by passing a detailed law prohibiting or allowing, for example, certain content moderation techniques. Instead, Congress left the courts to apply a simple set of general principles, some of which are drawn from prior common law ("good faith," for example). As a result, over the last 20-plus years, courts have been able to apply this statute to hundreds of cases involving a wide range of technologies, plaintiffs, defendants and actions involved.

Many Section 230 critics want Section 230 to look more like typical statute. They want Congress or the FCC to issue detailed requirements about who is to blame. This would undermine Section 230's greatest strength: its flexible and evolutionary common law-like nature.

Collective wisdom

Section 230's underlying ideas have proven sound, but its words are not sacred. There may be clearer ways to capture those underlying ideas. Moreover, because the statute sets out general principles that attempt to balance competing concerns about free speech, fostering innovation and limiting harmful speech, there will be times when the words don't seem to fit the situation.

It's unjust to hold the maker of a useful and beneficial tool responsible for how someone else misuses it, but it's also unjust when someone suffers serious harm and there is no one they can hold accountable. Courts must balance these concerns as they apply the words of Section 230. Maybe revisions to the statute could make it easier to find this balance.

No design methodology will produce good results if half the engineers think they are designing a fire extinguisher and the other half think they're designing a flame thrower.

Yet, ultimately, the concerns driving Section 230 reform today don't fit the criteria for a successful statute. This is a fast-evolving industry, and lawmakers fundamentally disagree on the problem itself.

To oversimplify, only slightly: Republican critics want to make platforms responsible for removing too much Republican speech while Democrat critics want to make platforms responsible for removing too little Republican speech. No design methodology will produce good results if half the engineers think they are designing a fire extinguisher and the other half think they're designing a flame thrower.

Even if legislators agreed on the goal, there is no universal solution to optimize content moderation because each platform has different needs and users. And the moderator's dilemma has only grown more serious as user-generated content has exploded. Prodigy processed 60,000 user posts per day in the mid-'90s. Today, Twitter processes that many tweets every 10 seconds.

Finally, courts are still incrementally evolving Section 230 case law as they consider other concerns that reformers would like to see resolved, such as whether and how third-party marketplaces are liable for the products that others sell.

Before lawmakers enact legislation to overwrite the collective wisdom of hundreds of court cases that applied Section 230 in the real world, they should be certain they understand what problem they seek to address and how changing Section 230 would affect it. As courts continue to develop Section 230 case law, Congress and regulators shouldn't disrupt them — especially because they don't agree on the problem or the solution. If they do, when the moderator's dilemma rears its head again, they'll be the ones who most deserve blame.