In the early Internet, it was impossible to find isolated references to people, places, and things. Many users navigated using directories like the original Yahoo, and early search tools like Gopher (structured documents) and Archie (FTP) were limited .
The search engines changed everything. Starting with Lycos and AltaVista, information was freed from obscurity. Suddenly, no matter where on the Internet your name might be mentioned, a search engine could find it.
On Monday, we discussed why the Internet is a new frontier. On Tuesday ,we questioned whether Section 230 of the Communications Decency Act of 1996 still fits the Internet of 2010. Today, I’ll explain how the rise of search engines since 1996 has changed reputation and privacy, and why CDA 230 subsidizes libel by preventing speakers and facilitators from internalizing the costs of their actions.
Google Has Changed How Information is Consumed
I don’t think Google is evil.
But Google is far from perfect. Google creates the illusion that just ten search results reflect some meaningful judgment on a person’s life. For example, the top five Google search results for any search term get 88% of the clicks. The over-attention given to the first few Google results is partly user error, but it’s also a form of rational ignorance on the part of searchers: Google gives good enough results most of the time, so there is little incentive to look deeper.
The attention given to the first few Google results would be fine if Google always provided accurate, balanced, and relevant information. Unfortunately, it doesn’t. Google has no way to measure whether websites contain information that is true, fair, or proportionate. Instead, Google uses rough heuristics—most notably the number of links to a page—to try to calculate a page’s popularity. Popularity substitutes for relevancy, often with comical results (remember “miserable failure?”).
Online, Google search drops users onto a website with no context or history of the site. There’s no indication whether a site is a parody (witness Salon being fooled by Landover Baptist), populated by anonymous trolls, a personal rant, or anything else. Of course, it is possible for users to perform this research for each and every site they visit–but the evidence is that they simply don’t (witness the Times of London being fooled by anonymous postings on a soccer website).
Rational ignorance? Possibly. Through experience, I’ve learned with law-related sites are reliable, which are tabloid, and which are garbage—but I haven’t had reason or opportunity to do the same for medicine, sports, fashion, or any of hundreds of other areas, and it’s questionable whether we should subsidize it further.
Google has Leveled the Playing Field – For Better and Worse
Google has elevated the ramblings of a lone speaker to the same visibility as the New York Times. This is a wonderful development for politics and freedom. It is a frightening development for personal privacy.
In the old days, the major media (think New York Times) was very unlikely to write about you. Your privacy was generally at the mercy of your neighbors and acquaintances—who often had to stake their own reputations when they chose to attack yours. If something rose to the level of defamation, it was usually easy to find the defendant and fight it out in court.
But today, anybody with a blog can (and all too often does) smear you, defame you, or invade your privacy. Their motivations are many: politics (if you read VC, you might have strong opinions), envy (think job promotion), mischief (think 4chan), etc. If you don’t have a big presence in Google before being attacked, Google will inevitably find the smear and bring it to the top of your search results: and tabloid material often rises to the top of a Google search because it gets the most clicks and attention.
This mechanism takes place even if the same content offline would undoubtedly be considered libelous or invasive-of-privacy. (If you disagree with offline liability for libel and invasion of privacy, you probably won’t agree with this either.) In many cases, you can’t find the original poster (it would take a lawyer, two subpoenas, and months). The host shrugs and says “CDA 230, not my problem” and rationally declines to name the creator (his customer).
Further, online defamation and privacy invasions can outlive the original speaker. In the offline world, most libelers stop once they are found. Online, hosts need not remove libelous or privacy-invasive information even after it has been found to be liability-creating. “Zombie content” lives on even after the original creator wants it gone (much to privacy advocates’ chagrin, Facebook does not delete all content when you delete your account), or even if the original creator has passed away or gone offline.
This has real consequences for real people. Consider false-but-hard-to-disprove allegations. How do you respond if a political opponent, a personal enemy, or simply a random stranger creates a blog claiming that you harassed or had an affair with a subordinate? What do people think when they see that in the first three Google results? It’s true that more speech can help push the false information down in search results, but it is near-impossible to prove the negative. And once that seed of doubt is planted (“did Obama shake hands with the President of Iran?” “was Kerry at a rally with Jane Fonda?” ) your name is forever tarnished.
Or consider the case of true-but-private information. Some anti-libertarians may question the Fourth Amendment by asking “If you have nothing to hide, then why should you care if we search you?” But should we cede our privacy so easily? What if a “peeping tom” photo of you ends up at the top of a search for your name? Your daughter’s name? It may be clearly illegal offline, but that doesn’t stop it from being distributed online without recourse.
CDA 230 Removes the Internalization of the Cost of Libel
I agree with commenters who have pointed out that stopping free speech online is (1) impractical, (2) inconsistent with the First Amendment, and (3) a bad idea.
But, consider how CDA 230 is subsidizing libel. Speech liability (libel, slander, invasion of privacy, etc) exists to make sure that a speaker can’t impose certain forms of harm on others (unfairly ruining a reputation) without feeling some cost herself. In the offline world, this risk of speech liability is largely internalized by the speaker and knowing facilitators: would-be authors of libelous publications know that they will be found and sued, and newspapers stop running advertisements they know to be false in order to stay out of court.
But online, the speakers often disappear thanks to anonymity and the lack of effective record-keeping by hosts. And, thanks to CDA 230, hosts suffer no liability even if they know that users are using the platform to defame others, profit from the resulting tabloid attention. CDA 230 goes so far that, under current law, a site owner could knowingly create a site that expressly encourages users to create false and malicious information.
In these cases, CDA 230 acts as a subsidy by removing liability (cost internalization) away from the speaker and host. There’s no incentive for hosts to keep records about their users; in a race to attract users, hosts have rationally advertised their lack of record-keeping–even though the lack of records imposes an external cost on defamation and privacy victims. There’s no incentive for hosts to remove content; Google rewards them (with web traffic) for keeping libelous material online–even though the material imposes external costs on victims. And there’s no incentive for users to not create libelous materials; in many cases there’s little practical chance of being found—even though it can take a victim years to clean up the damage.
The result has been a high-speed race to the tabloid bottom online among many content hosts. In a race to stand out in Google’s search results, which deliver users without context or background, some sites have encouraged tabloid anonymity (think JuicyCampus) rather than thought-through content (think VC). Maybe the market will correct the imbalance, but so far it has responded to the subsidy for libelous speech by producing more of it.
Tomorrow: Fixing the CDA 230 subsidy
David Thompson is co-author of the leading Internet policy book of 2010, Wild West 2.0 (Amazon) and general counsel of ReputationDefender, Inc.. The standard disclaimer applies: the views represented here are his alone and not those of any current or former employer.