« Thinking Like an Entrepreneur or a Lawyer? | Main | Haberman et al on Madoff »

Friday, July 03, 2009

Limiting Online Provider Immunity

My last post argued that Section 230 should be reinterpreted to immunize only content that comes from third parties via the interactive service offered by providers.  This would clear up many of the conflicts in the law, though providers would still be immune from some reprehensible conduct of its users.

Some types of conduct, however, should not be immunized even without reinterpreting the statute.  This final post in the series will tackle a couple of examples: marketing statements and celebrity impersonation.

Marketing Statements

As Eric Goldman points out, there is apparently some confusion about marketing statements made by companies that are rendered false by users. Two examples - First, an auction site might claim that its auctions are safe when in fact users commit fraud (court held no immunity).  Second, an adult web site might claim that its users are 18 when in fact some are minors who lie about their age (court held immunity because it was a "regurgitation" of user content).

The answer lies in a straightforward reading of the statute: providers shall not be held liable as speakers of third party content.  The statute says nothing about providers who make their own representations about what the third party content.  Providers who say "our users are polite" cannot be immune if the users are impolite.  Thus, neither the auction site nor the adult web site above would be immune.

This reading of the statute makes policy sense as well. While we want to encourage third party content and open contributions, we do not want to encourage false representations about such content. True, it is costly for providers to verify that their factual assertions are accurate, but not nearly as costly as trying to fact check each and every piece of content as it comes in.  We want providers who make representations about their quality of service to ensure that those representations are true, and letting a web site say "our users do ____" without actually enforcing a policy to make that fact true is bad policy.

Note that this provider/third party distinction would still immunize aggregation of third party content.  In Gentry, for example, eBay was immune even though it re-packaged the content in a variety of ways and aggregated user feedback. The difference is that eBay's interactive service managed the content, and eBay never represented that any of the information was true, verified, or otherwise accurate.

Celebrity Impersonation

Bill McGeveran discusses the problem of celebrity impersonation in online services such as Twitter. As he points out, lawsuits against the provider "would all be blocked, quite routinely, by section 230." But there is no reason why this must be so. 

First, celebrity impersonation is a violation of a right of publicity, which could be considered an intellectual property right (albeit a state right), which is exempted from Section 230's immunity.  As the comments to Prof. McGeveran's post make clear, there is a split (and more confusion) about whether rights of publicity are exempted.

Prof. McGeveran wonders whether it is good policy to give "more" protection to celebrities in this context when we give them less protection in other contexts, such as defamation. My answer is that ship has sailed.  We already give more protection to celebrities off-line, so doing so online, where you cannot actually see the celebrity's face or hear her voice makes distinguishing truth from fiction even more difficult. I also think that more harm than good befalls readers of celebrity impersonation (unless parody is clear), so the protection helps more than just the celebrity.

Even without considering a carve-out for rights of publicity, a straightforward statutory interpretation might still limit immunity for impersonating another person.  The statute immunizes a provider for content provided by a third party. This certainly doesn't immunize advertisements that a celebrity uses a service (see marketing representations, above), and it arguably does not immunize a provider's allowing user to impersonate another.  Yes, the provider would be immune from any of the content posted about the third party, but not necessarily for providing a system that does not, for example, check identity.  Identity is arguably not content.  This, I realize, is a much more controversial reading of the statute, but it arguably could apply even to non-celebrities.  That said, I think the policy implications of such a reading are troublesome, as the cost of verifying identity of all users - both in verification and in loss of anonymity - is not one we want providers to bear.

In any event, loss of immunity does not mean that the provider is necessarily liable.  I suspect that a common law notice and takedown procedure would work just fine for celebrity impersonations.  It is relatively easy to prove that a user is not in fact a celebrity, and the complainer is, in fact, a celebrity.

Section 230 is a powerful ally for online providers and a good thing for the internet generally.  These last two posts, however, show that courts should more closely read and follow the language of the statute, and that they can do so in a way that prevents at least some of the harms that come with unwanted user content without sacrificing the good that Section 230 provides.

This is not only my last post in this series, but my last post in this guest blogging stint.  Thanks for having me; I had a great time, and hope to be back soon!

Posted by Michael Risch on July 3, 2009 at 08:48 AM in Legal Theory | Permalink


TrackBack URL for this entry:

Listed below are links to weblogs that reference Limiting Online Provider Immunity:


Mickey Kaus has evidently just discovered Section 230 - it's fun to get a reaction from someone new to the statute:

Posted by: Derek Bambauer | Jul 7, 2009 11:28:34 AM

Thanks for the thoughtful reply, Bill. Sorry if I misread some of your post. My point about the right of publicity is that celebrities don't get as much protection off-line for defamation, etc., and so they get the IP protection.

You are right that non-celebrities get no such tort relief while celebrities do. I also agree that there can be disagreement about whether it's time to call the ship back in on that.

Posted by: Michael Risch | Jul 6, 2009 3:26:16 PM


Thanks again for your thoughtful and interesting series of posts about section 230. But I think this post here misread my earlier response.

My post about the impersonation of Tony LaRussa on Twitter explicitly agreed with your statement that a right of publicity might, and a trademark claim certainly would, fall under the IP exception to section 230 immunity. (The full quotes from my post are: "LaRussa’s actual grievances sound like they should give rise to defamation or false light, or perhaps the appropriation tort. But these would all be blocked, quite routinely, by section 230. ... [T]his case highlights the possibility of a loophole for celebrities who can recast privacy-like claims under trademark law (and possibly also rights of publicity, if those are interpreted as intellectual property under the language of section 230(d)).")

More generally, I think you are missing how 230 can go a step further than we already see in offline publicity rights toward elevating celebrities' legal remedies above those available to others. Suppose the dead-tree newspaper published something over my name written by an impostor. I do not have much of a right of publicity claim against the paper because my identity lacks commercial value, but I still have other torts like appropriation and possibly defamation, false light, or even infliction of emotional distress depending on the circumstances. Moved online, Tony LaRussa keeps his IP claims, but I lose my garden-variety tort claims under section 230. (Of course, at least in principle everyone keeps their claims against the actual impostors in both cases, so set that aside.) The key is that celebrities have IP rights in their identities. Even though makes sense in some contexts, I think here it leads to perverse results.

LaRussa's lawsuit may be sui generis and not indicate a wider problem. (It also relies on what I consider a bogus premise, that his name on a Twitter account acts as an endorsement of Twitter.) But I think it is actually a nice window into some perils of propertizing celebrity identity, whether through publicity or trademark rights. To the degree that "ship has sailed," well, we need to call it back toward the shore.

Finally, to be clear, I have written before that some form of immunity like section 230 is absolutely essential for a robust internet. As your series of posts illuminates, however, designing one that works properly and clearly is not always easy. (Indeed, Derek Bambauer and I are disagreeing on our blog right now about another issue of section 230 interpretation!)

- Bill

Posted by: William McGeveran | Jul 6, 2009 3:01:50 PM

The comments to this entry are closed.