December 3, 2005
Okay, I am at the Regulating Search conference at Yale, and will post notes in this entry as the day progresses.
Just got done with the first panel, which I was on.
Andrei Broder from Yahoo has a neat high-level outline: Search is transforming from syntactic (e.g. matching keywords to text on a page) to semantic (e.g. understanding what it is you’re actually talking about), and will continue on to “information supply” (e.g. no explicit searching — information just appears as you need it).
Now, the panel on regulation.
Barbara van Schewick drops an interesting factoid: In terms of eventual transactions, there’s a drop-off of more than fifty percent from the first search result to the second! Wow.
Renata Hesse, from the antitrust section of the Department of Justice, responds briefly to the idea of Google FOIA… with horror. Just because it has the potential to create so much work! (Conference attendee Michael Zimmer thinks it’s interesting, though.)
In response to a question, Yahoo’s Andrei Broder distinguishes between “navigational searching” (e.g. I am looking for the University of Chicago law faculty blog, but don’t know the URL) ) and “informational searching” (e.g. I am looking for a good constitutional lawyer). Apparently about 25% of all searches are the former.
Lunch break! And the entry break as well. Read on.
Back for panel three.
Martin Schwimmer worked on the very first case involving a domain name dispute. Cool.
Jon Baumgarten is a copyright lawyer. He says, when it comes to copyright, there’s a difference between a “good use” and a “fair use.” Making everything accessible and searchable is a good use, he agrees — it would be handy and socially productive. But that doesn’t mean it’s a fair, or permissible, use.
He sees Google’s book-scanning as a “dramatic truncation” of copyright law — so that it encompasses distribution (e.g. showing it on a screen) but not reproduction (e.g. keeping it on a hard drive) rights.
Ooh! Now Paul Aiken from the Authors Guild (the group suing Google) is speaking!
Here’s the story he tells: Google gets permission to do its book-scanning. Certainly Yahoo and Microsoft would follow suite. So it becomes a general search-engine exemption. But then… anybody can create a search engine. So why wouldn’t I just scan some books I like and make a little search engine of my own? Perfectly legal, right?
So maybe Google can secure its scanned library. But can I? What happens when these scanned books start slipping out on the internet? And there’s a rad ebook reader to view them on? Basically it’s the specter of the MP3-ization of books.
Jason Schultz from the EFF is now cutting straight to the heart of things: the idea of intellectual property itself. He says IP is not the same as owning land. IP is a government grant, and with it comes obligations. All IP rights must exist in the context of the public good. His position paper is worth reading.
He cautions against pre-emptively outlawing stuff like Google’s book-scanning, because if we do, we’ll never know what the results really are. As policymakers, we need data, and we don’t get data unless we experiment. If it turns into a huge mess, we can always make new rules and even compensate people retroactively.
Daphne Keller of Google returns to first principles of IP as well. She cites “The Creation of the Media” by Paul Starr, which traces the U.S. government’s long history of subsidizing the broadest possible distribution of information.
There’s some good debate over Google’s book-scanning. I am seeing both sides of the story here (though still come down on the side of Google).
Time for panel four, baby!
Did you know that anonymous speech is protected by the First Amendment? I didn’t — Aden Fine from the ACLU says it is. Does that mean anonymous searching is or should be protected, too?
Mike Godwin (coiner of Godwin’s law!) is talking about the instrumental purpose of copyright — its larger mission — and saying it is not the same as property. Authors’ rights, he says, are not ends in and of themselves.
He responds to Baumgarten’s point (above): He says even if good uses are not automatically fair uses, goodness of use ought to factor into the equation. What is policy for, otherwise?
Another point: Is it reasonable to talk about copyright in an era when the incremental cost of making copies is rapidly approaching zero? Is there some other framework that would be more appropriate?
Now Chris Hoofnagle from the real EPIC is talking about “commercial data brokers” — companies like Choicepoint — that collect personal information from public records and then use it for, you know, evil.
Also, Hoofnagle says we’re too easy on Google — we’re not having a full debate about the privacy implications of all that Google does, and all it could do. “Don’t be evil” is not a real policy, he says.
Good line: “When people say ‘Google’, we think ‘Tech Jesus.’”
Now, Alan Davidson from Google’s Washington office is emphasizing that policies need to operate at scale — huge scale — because that’s where search engines live.
It’s wrapping up now… cool stuff!