Sunday, December 09, 2007

Reputation economies, Panel IV: Ownership of Cyber-Reputation

Moderator: Eddan Katz

John Clippinger: He’ll play Polyanna. Tech is facilitating rapid learning about the consequences of lack of privacy, as we see with Facebook. The internet is/should be edge-oriented, with the locus of control in the individual. He has been working on a framework for interoperable and separable identities. Then people can aggregate info from profiles into a metanetwork. The project – “Higgins” – is making its way into the commercial sphere.

Commercial entities fear privacy breaches – they don’t want to be targets of hackers. The vendor wants to know what you want, not who you are. If it doesn’t have to store your address/other info, why would it?

Eric Goldman: As general counsel of ePinions, faced many of these issues years ago. Here he focuses on one problem: ePinions needed a product catalog, which means that products needed to be identified uniquely and persistently, just like people.

There is an uncomfortable interplay between IP and reputation systems. IP can be used to suppress the flow of important and valuable information. Herein of taxonomies: classification schemes, to direct commentary at the appropriate product and direct searches to the right results. IP can hinder this if people claim proprietary rights in product catalogs. Catalog provision is an active business, but inefficient – some of which is due to IP.

Cases claiming copyright in part numbers attached to products have been attempts to suppress competition by preventing ads for interoperability; they also create risks for those publishing a catalog. So far, the courts have rejected such claims, but the issue is far from over. The ADA case allows rights over a catalog to control secondary offerings.

Even if there’s no copyright in a single number, other IP rights can still be asserted: if you scoop it from the plaintiff’s servers, there’s trespass to chattels, contract, robots.txt, CFAA, and related state law claims. Claiming rights over catalogs (1) requires wasteful duplication of investment and (2) creates different catalogs, making it harder for consumers to map from A to B and creating isolated islands of consumer opinion.

Solution: government involvement! There should be nonexcludable freely available product catalogs: the ISBN as the flagship example.

Another solution: sites like ePinions are not entitled to trademark’s first sale doctrine because they’re not selling the goods. They need clear rights to host reviews and offer a taxonomy even though they’re commercial entities, otherwise TM owners will cherry-pick and get rid of negative reviews. Possibilities: (1) categorical safe harbor; (2) innocent printer/publisher exception (that doesn’t stop cherry-picking though, just leaves the provider subject to notice and takedown); (3) categorical exemption for commercial referential trademark uses.

Question from Frank Pasquale: he worries about distorted search engines/databases. If one is sponsored by Sony, should Panasonic get to put an asterisk/link on its results?

Goldman: He won’t agree to that. But if you did want to do it, a universal ID would make that easier.

Sutor: RFID is making this a real issue right now.

Bob Sutor: The number of reputation formulas is unbounded, but we can pick contexts. If we’re lucky, we can get it down to 5-8 ways of representing people.

Standards become important when information leaves one provider’s domain or application – inside, it doesn’t matter. So standards and privacy are inherently intertwined. The problems are not from what information is available but from who can do what with it. Unfortunately, security tends to be the last thing people think of – build the neat tools first, then worry about privacy later.

There are two extremes: hyper-structure, as with the semantic web, or PageRank/informal tagging (which tends to work pretty well). People have not generally marked up web pages systematically, but we do have a lot of information about people and objects, whether volunteered in profiles or inferred/computed. This means the semantic web might start to be relevant.

Mozelle Thompson: Who gets to set the standards, under what conditions? Market participants will seize on opportunities to game the system.

Consumer perspective: Ownership is irrelevant; control is key.

Facebook gives a nod to offline connections. Proxies: people view their data as representations of themselves. People consider security as an issue of how much risk a company is willing to expose them – its customers – to.

Privacy is not binary. It’s not that more information is bad, less information is good. People are becoming more strategic in their uses. They expect sites to tailor themselves to specific users, and thus presume that the site has access to some information. It’s not a top down relationship but horizontal – people are disclosing information to reach other people, not for the benefit of the marketers.

Government’s task: give people management tools. Neither government nor business has educated people on how to use tools and how to use judgment, not just knowledge.

European regulation emphasizes self-determination – the ability to withhold information -- but this is difficult when people voluntarily provide info to social networking services. At the same time hands-off US regulators are beginning to recognize that people may lack sufficient information to make rational choices. There is thus a de facto convergence between the US and EU approaches.

Rebecca Tushnet: If you were to look at my Blogger profile (not that there’s reason to do so) you’d see a picture of me. But on my personal journal, my user icon is an image of Wonder Woman. Why? (Why not?) It says something about who I am, or who I want to be, online. We identify ourselves to ourselves and others by what we consume: favorite books, movies, brands, cars – all the Apple and Dell logos I see by looking out into the audience. Like my Margaux Lange pin, we are all made out of others’ faces.

This is reduced to its most abstract form in online interest lists/profiles, as shown on social networking sites. LiveJournal, with which I am most familiar, has implemented interest filtering – it can stop you from searching for particular interests. This has implications for ownership of reputations – IP owners could assert rights to control who gets to announce affiliation with them. Attempts to control references to TMs, or uses of copyrighted works as part of a person’s self-definition, can threaten established user reputations. A DMCA notice could literally take my face away from me.

Something that hasn’t been said about the Star Wars Kid at this conference: the interested party we haven’t heard about is Lucasfilm. The Star Wars Kid, and the remix videos starring him, affect the brand’s reputation as well. And Lucasfilm would probably have a better chance than the Kid of getting those remixes taken down!

Reputations, that is, raise the same IP problems as user-generated content generally does. Reputations can be based largely on use of others’ IP.

Examples (and I should say that these are differentially vulnerable to IP claims): Linking – popular bloggers can become popular by being good at sorting and identifying others’ works. This is great in that editors finally get some respect, of a sort. But copyright owners have been willing to claim rights against linking.

Fan videos/mashups: the creators get credit for their creative editing skills. But they’re hostage to DMCA notices – if a video is taken down, the poster loses all the comments, ratings, and other reputational benefits, and even if it pops back up again elsewhere the accumulated “merit” ratings from others are lost.

Likewise, eBay sellers can lose their accounts, including years of feedback, if they’re reported for selling counterfeit goods. If they are selling counterfeits, that’s fine, but there are manufacturers who take the position that any resale is unauthorized, and that’s a problem.

Benkler and others see blogs and other forms of user-generated content as the capillaries of the internet. IP claims have the potential to work as a tourniquet, cutting off the upward flow of information. (Related to this is the question of whether everyone on the internet wants to be part of a greater body. Maybe I don’t want to be a finger; maybe I want to stay in a subcommunity. This might be especially likely if my subcommunity is exposed to IP claims if it does become more visible.)

Congratulations to the organizers on an interesting and provocative conference. I’m not sure the “ownership” panel really came together, but the parts were pretty interesting.

No comments: