With some luck, the column will ripple across the Internet, and its insights (assuming there will be any; I'm just getting warmed up here) will be repeated by bloggers and news aggregators. But I have no way tracing those ripples, short of doing a Technorati or Google search on each phrase that's worth quoting. Admittedly, that wouldn't take long in my case, but pretend for a moment I'm Steve Lopez or Ron Brownstein, whose columns leave a considerable electronic wake. What would I gain from knowing where my bons mots have reappeared, and what would I do with that knowledge?
These are the questions raised by Attributor, a Silicon Valley start-up that is developing a way to trace text, video and other media as it gets posted and reposted around the Net. The company, which has yet to offer services publicly, is one of more than a dozen firms offering ways to identify content online automaticallytypically by creating what amounts to a digital fingerprint of each of their clients' products, then storing the fingerprints in a database.
Many of Attributor's competitors position themselves as gatekeepers. As a file is uploaded to a site such as YouTube or traded across a file-sharing network such as Kazaa, their technology will extract its fingerprint and compare it against their databases. If the fingerprints match, they block the transmission or demand payment. Attributor takes a different tack, combing through everything published on the Web for items that match their clients' wares. Unlike Audible Magic, whose filters are used by the iMesh file-sharing software (and, reportedly, YouTube), Attributor won't be in position to stop something from being posted or shared. Instead, it's primarily a tool for copyright holders to know where and how their work is being reproduced.
In theory, this is a much more attractive and Net-friendly approach than trying to bottle up content online. The more a work spreads across the Web, the larger the potential audience. And for most digital goods, the larger the audience, the bigger the potential payoff (the exception being things like trade secrets, whose value diminishes as they spread). Yet the pervasive scrutiny enabled by Attributor carries a whiff of menace, as it sets the stage for an escalating fight over speech and fair use online. How this plays out will depend largely on what copyright owners do with the information Attributor hands them.
Let's break this down into three scenarios.
The first, and easiest conceptually, is when a site posts an entire work (or a segment that can stand alone, such as a "Saturday Night Live" skit). Under the 1998 Digital Millennium Copyright Act, copyright holders can force the work to be taken down, but that's no way to add revenue. Instead, if the site is selling ads, Attributor wants to help its customers collect a piece of the action associated with their works. For example, if a news site were to post a copy of this column and ring it with ads for DVD-copying software, Attributor might give the site a choice: pay me a share of the ad revenue, pay me a flat fee, or take my column down. (Attributor's specific plans for dealing with sites that post its clients' works are as yet undisclosed.) That seems like an appropriate choice, given that the site was making money off of a column it didn't own or license.
The second scenario is when a site uses just an excerpt. In this case, the site is more likely to argue that it has made fair use of the work and doesn't need to pay the copyright owner. Jim Brock, chief executive officer of Attributor, said in a recent interview that his firm's approach should help customers judge whether a use is fair or not by revealing the context in which the excerpt was used, and how much of the original work was posted. Still, as Brock notes, copyright owners want to draw much tighter boundaries around fair use than advertiser-supported websites do. Meanwhile, Congress has set only general guidelines for fair use, leaving the courts to resolve disputes on a case-by-case basis.
The third scenario is when a site blends the work with other content to create something new. The fair-use questions here are even tougher than in the previous example, in part because digital technology makes it easier than ever before to manipulate, mash up and remix content. Brock argued that standards would emerge for judging these creations, but at this point the waters are largely uncharted.
Beyond that, there's the issue of what to do when a work pops up on non-commercial sites or those that aren't really making money. These are the sites that can break a content owner's business model. After all, Internet users are adept at finding free outlets for the goods they covet. An Attributor client that simply wants to be credited for its work can live in harmony with such sites, but what about one that supports its business with advertising?
The emergence of Attributor and other content-identification tools is emblematic of the gradual transformation of the Internet from a free-wheeling communications pipeline into a more rule-bound platform for commerce. There are potential benefits for everybody; it's hard for companies to make their products available free without being able to tell advertisers where it runs and how large the audience is. So it's conceivable that Attributor could coax some of the magazine and book publishers who have been leery of the Net to make more of their works available online. Still, the transformation has some trade-offs, including greater scrutiny of what people say and do on the Web and an intensifying battle over how to define fair use.
Jon Healey is a Times editorial writer.