SIMS 141 - Intellectual Property & Search: Jason Schultz,...

Uploaded by UCBerkeley on 20.08.2007

OK. So basically I'm going to talk about three questions.
I'm going to focus primarily on the first two, and the
third one's just going to hang there at the end.
Mostly so that we can talk about it if have a
little time at the end.
But the first is, who controls our access to
information online?
And what mechanisms are guiding us as we search
around trying to find things.
The second is, should our access to information online
be more, or less, or equal to that that we can find offline.
Should it make any difference that we're doing it
digitally over a network or something like that?
And then, third, after I talk about the issues with
intellectual property and search, what are the kind of
competing public policies that are worth thinking about when
trying to decide what the right rules should be.
Where should property rights control, and where should
other things control. And how should we
balance all that.

OK. So the first thing is, who controls our access
to information online. Well, so, as you can guess I
will be talking a bit about intellectual property rights
owners and how they can control what we can
find and what we can't.
And what we can download and what we can't.
And what we shouldn't download but do anyway,
and things like that.

The second is that there are the providers.
And so in the search context I just listed a couple.
But they also obviously, and Hal's talk showed you, control
not only the search results but other types of information
through ads and another things that might come to us,
that we can access online.
They may not be the only source, obviously, they're
competing with each other and there are other sources,
but they're certainly one.
There's this guy. And the people who work for him.
Who also have some control over the information
we access online.

There's the court system.
And we'll talk a little bit more about that, about when the
courts get involved and shut down your access to information.
There's the market.
This is the only graph I still understand from economics.

And that's a barely.

Sorry, Hal.

And then there's France.

We'll talk about that, too. About how other countries in
the global, international, environment's
controlling things. And then there's you. Well, some
approximation of you. And more probably an
approximation of me.

But there's the user, who also controls what
they want to find.
Sometimes inadvertently online as well.
And so, the reality is that all of these things
come into play, right? So mostly I'm going to
talk about intellectual property rights here.
But it's useful to keep in context that all these things
are sort of interplaying and interacting to give us
information as we go online.

So I'm actually going to skip through most of this.
Because so many of you raised your hands.
But intellectual property is basically, when you dumb it
down, patents, copyrights, trademarks and trade secrets.
There are some other things.
But patents are basically a sort of property
right on an idea. A way to do something, or
something you create that's never been done before.
Formulas, chemical creations, products, things like that.

Copyright covers expressions of ideas.
That's one way to phrase it, sort of creativity
that comes forth. So it's not really the idea
of boy meets girl but the implementation of
boy meets girl.
In a film or a music or video game or books or a web page.
Those are all copyrighted works.

Trademarks are basically brand names, as most of you know.
And then trade secrets are secret formulas, or
how to, like a special way to do something.
So, one of the reasons that I'm sure Hal did not tell you the
exact way in which Google computes many of the
things he talked about.
Is because that is a trade secret.
And that if he were to disclose it and the world were to know,
then Google may not have the competitive advantage
that it thinks it has.

The most famous example is the Coke formula, actually.
The formula for Coke, that apparently only six people in
the world or something know it.
And they have cyanide capsules in their brains that go off if
they ever try and disclose it, something like that.

So, just for a second, looking at it from the perspective of
the search providers themselves.
Like Google, Yahoo, A-9, AltaVista, all these companies.
The primary way in which they have defended their turf,
through the intellectual property, has been this
combination of trade secret, patent, and trademark.
So they start out usually with some secret formula, some
secret sauce, of how they come up with results.
How relevant they are, things like that.
They end up over time, usually patenting them.
And one of the parts of the patent process is,
you have to disclose it. You have to file it with the
patent office and then now after 18 months, typically
that gets published. And then eventually your
patent gets issued, and that's published.
And so the whole world will then learn, to some
degree, how you do it.
But then, of course, you'll have a property right to
keep people out of it.
And, then the third part, though, is trademark.
And you've seen this, and Google's actually been seen
as a brand name, really.
In that if Google attaches its name, its G, to something and
it becomes G-money right?
So, it's pretty much that the strength of the brand
associated with these kind of of secrets and patents that
keeps them kind of invested in intellectual property.
On the other side, though, and this is primarily what I'm
going to talk about today.
Is what are the threats to people trying to develop
search technology. One of the things that the
Electronic Frontier Foundation is concerned with is, the sort
of climate for innovation.
Sort of in an innovation ecosystem.
How easy is it for new ideas to get implemented, new
technologies to come about.
New, exciting, things to happen in the world of technology,
in the online world. And one of the things we're
finding more and more is that but for intellectual property
rights, we would have a whole lot of things that we
would want but can't have.
And sometimes that's a good decision.
Sometimes it actually is the right decision to hold back
on innovation because there's some huge social cost.
But more often than not, we tend to feel that the
intellectual property rights owners tend to overstep a
little and try and inhibit innovation because they're
trying to hold onto a market that they used to have and is
slowly slipping away from them.
That's sort of our philosophy. So, I'll disclose my bias
right now, as to that.
But you're seeing that happen a lot in search.
And I'm going to talk about the various instances.

So, first things first, most of you raised your hand, you
knew what trademarks were.
The two main policies behind the idea of a trademark, why
we have it, why the government grants them, and why people
invest in them, are that it promotes investment in
a brand name or logo.
And then it also prevents consumer confusion.
That's really key, as I'm going to talk about the legal cases.
They almost always center around this
notion of confusion.
The idea is that if you went to the supermarket and there were
four brands of Tide detergent, one which was the real Tide and
three of which were knockoffs that were half as good
and almost as much, you might buy the wrong one.
You would be confused, and you'd end up not getting
what you wanted. It would take a sale away from
the real, true Tide, and that would be bad for everyone.
So, by giving people rights in a trademark, they are ah
allowed to go after the knockoffs and sue them, shut
them down, get them out of the supermarket, get them off
the streets hopefully. And therefore their brand will
be the only one that really symbolizes who they are.
It's a source identifier.
That's how that's considered. So Tide, when you see it in
the store, means a particular company's product.
And no-one else. So it's very much an
information location tool in a sort of rudimentary sense.
That it's guiding you to what you think you want.

But also that there is this notion.
And the thing is that trademark in some ways is not even really
a property right so much as a consumer protection law.
But a lot of people treat it like a property right.
They treat it like, well I registered this
trademark, now I own it.
And even if consumers aren't being confused, someone else
wants to do something with it, it's mine.
It's not theirs. I don't share well.
I'm going to sue them. Right?
So that's kind of what we'll be talking about which goes on in
the search space, is that things are getting blurry and
a lot of trademark owners are getting really uppity.

Oh, by the way this is my favorite example of
what's going on here. Because this is one place
where Google has been sued a bunch of times.
But Google has actually sued for confusing
people about which is the discounts web page
to go to for items.
And I actually think that in this instance, Google is
showing where trademark can be, in some circumstances,
appropriately used. Because you can at least
see that there's a kind of confusion that
could happen, right?
Outside of context, some people could go to one site or the
other and not know which one.
Although I think that by now the distinctive Google colors
are pretty much out there.
But it's a closer case, and certainly one that wasn't
totally bogus to bring.

OK. So, I'm glad that Hal gave a bunch of basics
on the keyword stuff.
Because I'm going to gloss over it now.
But, basically, as you know, that he said, this keyword
advertising search based idea is that you would
type in some phrase.
And ads will come up around the search results.
Well, the big conflict that's come up around trademark, and
I'll talk about copyright later, I'm going to focus on
trademark just for the next ten minutes or so, is around what
happens if you type in a trademark.
What happens if you type in Apple Computer, what happens
if you type in Coca-Cola, or in this case 1-800 Contacts.

And I'd use this because 1-800 Contacts has been
particularly litigious.
They've sued a lot of people including, at one point,
threatening Google. I don't actually
remember if they sued. .
Or they settled. But they're one of the
people who goes after people. And as you'll notice, they
actually have the promoted ad.
This is from today, or whatever, so it may
change at any moment.
But they have their promoted ad.
But if you notice over here, there's a and
there's a, and I don't know whether those are
competitors, but let's assume for the sake
of argument they are. Let's say these are competitors
who have been on that keyword so that they have ads that are
tied to the keyword, which is also the trademark.

So there's this sort of weird sense of, OK, well but if it's
1-800 Contacts' trademark, maybe they should be the
only ones who have it. But over here there
are these competitors. And everyone's bidding on it.
And we'll sort through that. But that's the general
conflict that's going on. Is, what we do about this
kind of a situation.

One other thing I want to say, and I'll give some examples to
help make this more concrete, is that there are different
types of trademarks, too. And that helps keep this
kind of clear in your head.
Is that there are some trademarks, like Exxon, that
don't actually mean anything in the English language.
It's a made-up word, it's what's called arbitrary
in trademark laws. An arbitrary mark.
There's things like Coppertone which is what's called a
suggestive mark because it's again not a real word.
But it sort of suggests tanning of the skin.
It sort of brings to mind a certain idea.
And then there's descriptive marks.
And that, so I use Omaha Steaks and Mrs Field's
Cookies as examples.
Where they are trademarks, but they are also descriptive of
something that could not be that thing.
So if I am an independent steak provider and happen to live in
Omaha, I might have an interest in describing my steaks
as Omaha steaks. Right?
I might want to bid on Omaha Steaks at Google as a keyboard
even though I don't want to infringe the Omaha
Steaks trademark. I don't want to go there.
But it in fact an accurate example, an accurate
description of things that have to do with me and my business.
So that you have this crossover between something that is
serving as a trademark and something that's quite
an accurate description.
I like Mrs Field's example, also, because that's my
mother's maiden name. So I always had the vision of
my grandmother making cookies and putting them out
and selling them. And I could call those
Mrs. Field's cookies.
And then what if I took out the keywords for that on
Google and had pictures of my grandma and things.
And there's some legitimate interest I have in that.
And so this is what intellectual property law is
supposed to do is trying to figure out a balance
between these two things.
Protecting the legitimate rights of trademark holders
stake, so that they invest in trademarks and they
serve their purpose.
But not precluding everyone.
Excuse me, not precluding everyone else from using
them for legitimate purposes.

The last thing I want to mention is that there are
these things called generic trademarks.
Or trademark that become generic.
And Aspirin is actually a good example of that.
Aspirin used to be a trademark but when there's a certain
thing, when people tend to use the word over and over and
over again, it suffers from what they call genericide.
Where it becomes part of the English language.
So actually Escalator used to be a trademark.
Elevator used to be a trademark.
Thermos used to be a trademark.
For a while Xerox and Kleenex were kind of on the edge.
Depending on you talk to, they've kind of
pulled back or not. And in fact.
Google is concerned about this as well.
That their own trademark will become synonymous with search,
and therefore they'll actually lose it to some degree.
That's why they emphasize that to Google something means to
actually search on Google as opposed to Yahoo or something.
But, anyway, so it's kind of funny.
So Aspirin used to be a brand name for this kind of thing.

So I'm going to give an example in the real world about what's
going on with trademarks that is this conceptual framework
for how people are starting to talk about keyword searching.
So this is going to happen in the real world.
And then we're going to say, OK, is this true.
That this is what's happening in the online world as well.
So you're a consumer, right?
You're hungry.
You're in your car and you're driving on the freeway.
And you're driving along and you see a road sign
for Super Duper Weenie.
And you say, wow, that sounds pretty good.
Maybe if I exit here I'll actually find Super Deeper
Weenie and go eat there.
And then if I give them my cash, I'll get some food and
I will no longer be hungry.

So, but Super Duper Weenie has a competitor.
The Evil Lentini's who runs a pizza and so-called
family restaurant.
And they say, wow, we really like those consumer dollars.
But we have kind of a dumpy place.

But wait a minute, and we can't really compete well with the
Super Duper Weenie, because they're all snazzy, right?
But maybe if we put up a Super Duper Weenie road sign by our
exit, people will come here.

Thinking they're going to Super Duper Weenie,
they'll be confused, but they're still hungry.
They still have cash.
Even though we're dumpy, they'll buy our food.

Because you're driving along, it's late at night, or you're
going somewhere and you get off the freeway because
you see a sign. And you get there, and you're
like, well, it's not what I was looking for but I'm here
anyway and I'm hungry and fine I'll buy your stupid pizza.
And so, I mean, you do this on campus all the time, right?
You know, I'm here, I'll buy it.
And so this is considered trademark infringement
by some courts. Under a theory called
initial interest confusion.

And so this is a theory that's talked about.
Which is that when you actually make the purchase you're
actually not confused. You're Evil Lentini's and
you're buying the stupid pizza, and you're OK at
that point as a consumer.
But you were kind of lured there under false
pretenses, right? It's not like the grocery
store where you're buying Tide and it's not really Tide.
You know that you're buying the other brand, but you wouldn't
have bought it if you hadn't initially been deceived.
So, courts have kind of, some courts in the US, not all of
them, have adopted this as a theory.
That trademark holders could sue other people who are using
their trademarks in a way that lures people away.
And the key here, as I mentioned, is that how
likely is it that consumers are actually being
confused by this. That they're being misled
in a really overt way.
And to some degree, how much of a pain in the butt is it
for them to get back on the freeway.
When you're driving around in a physical car and you're
super hungry, it might be kind of a pain.
But when you're on the internet and you can click the back
button, maybe not so much.
So, courts are trying to struggle with this.
So we get back to this question of, do keywords confuse people?

And that's a really good question.
And what's interesting is that you see from Hal's talk how
hard it is to compute all this information that's going on
just relating to which ad should get placed where.
And how many billions of searches are going
on in a given month.
How in the world do you figure out who's confused
on the internet. Well, we're all confused
to some degree, right?
Generally, about everything on the internet.
But how do you tell all these millions of consumers, all
over the country, all over the world, who is being misled.
And who's not? Well, some things to think about.
So I should say that some courts have just said,
well, we're just going to assume everyone confused.
And this is just evil and bad because we don't see
any good reason in it.
And probably the only reason people are doing it is
nefarious reasons, right?
But that's not necessarily the case.
There's some competing policy reasons that make
a lot of sense for this. And it's a question
of analogies. Oftentimes when you get into
the online world and the legal world and court battles, it's
a battle of who has the best analogy.
Is it more like a toaster or a TV.

So, the question is, are we being diverted or
are we being informed?
Well, if you go to the grocery store you'll often see brands
lined up next to each other.
And in fact in most grocery store chains, this is a
little-known fact, you can pay for placement.
Next to big brand names.
So if you are a knockoff soda, you can pay to be placed on
the shelf right next to Coca-Cola and Pepsi.
And that's completely legitimate.
And no-one has ever brought an initial interest
confusion lawsuit over that.
Because someone walked up to the aisle and looks and says,
oh, there's Coke but here's something that's Cola and it's
50c less per can, and maybe I want that instead, right?
And this goes on all the time.
Same thing happens in Macy's and department stores actually.
Many department stores will have their own lesser brand.
So you don't get seven jeans but you have a sort of knockoff
Macy's jeans that are sort of looking like them but not quite.
But they're an different brand and they're half the price.
They'll often put those, it's actually funny, if you
look at the architecture of department stores.
As you walk back to the huge brand name section, the Calvin
Klein or whatever, along the way you will see the
knockoff brands.
Because they've strategically placed them there.
So as you're walking you're like, wait a minute, 50% off,
that's not so bad, right?

And that's all strategic, it's all business, it's all
the way things are done. Because in the end when you go
to the counter and the register you're not actually confused.
You're actually buying something that's useful to you.
You found a better deal.

Another example of this is in comparative advertising.
So here's this ad about three different minivans.
And apparently they're fixated on exactly whether it's 201
inches or 199, or whatever.
Apparently this makes a difference.
But it's using brand names.
Chevrolet, Dodge, and Ford.
But what if Ford had published this in a magazine.
And let's say for some reason they think the size matters
and 201 is a better inch length for them.
Well, Chevrolet could maybe try and sue them and say aha, well,
sure, Ford, you list all this information, but people
see Chevrolet first.
It's at the top of the page.
You are using our Chevrolet trademark to lure people, to
trick them to looking at the ads so that you can then try
and sell them Windstars.

So you sort of see that there are similarities in
the sort of logic here. Well, is it more like pulling
someone off the road to go eat at an evil pizza diner, or is
it more like comparative advertising or offering
people choices. And that's kind of what the
law is trying to sort out with this keyword problem.
So, if we go back to the Google page for 1-800 Contacts, it's
really the sort of core question is, is it pro-
or anti-competitive? Is it helping consumers, is it
helping markets, or is it hurting them, stealing things
away from someone who had value or is it actually providing
alternative options.
So obviously certain trademark holders feel it's
diluting their brand and stealing their value.
And the providers, Google and the others, generally
believe that it's in fact offering other options.
It's offering, it's like a phonebook or a card
catalog or something else. It's offering people options.
And they're trying to say, our users are actually not as
dumb as you think they are.
And this actually helps them.

So we have a case.
A lawsuit, that happened between GEICO and Google.
Which is now settled.
GEICO is the insurance company, car insurance company.
It's actually the government employee insurance
collective something. I didn't realize it was
a government employee program to start with.
And then it went private.
And they got upset because Google was selling
the keyword GEICO.
And tying it to ads, just as Hal laid out.
And so they sued.
And they actually had a trial.
And it was in Virginia, and I want to point out that this is
an example, actually, of what they call, in trademark,
a normative use.
So GEICO is a made-up word.
There is no, even though they have gecko as the mascot, it's
the word GEICO itself does not have an English meaning.
We'll talk in a second about a lawsuit where the words
actually had a dictionary meaning as well.
And, so GEICO sued and said, you can't ever have anyone
tied to our keyboard except us.
And they produced this big survey.
They did a consumer confusion survey where they showed
people a bunch of web pages in these crazy numbers.
And almost 70% of people were totally confused and thought,
oh my God I'm going to reach GEICO going through the
sponsored links on the side.
Even when they weren't GEICO links.
They must be affiliated, and 20% said that they would
purchase insurance from GEICO if they clicked through
on the sponsored links. But actually, Google's
attorneys did a pretty good job in debunking this survey and
showing it had mathematical and methological flaws.
And so the court actually doubted it.
And ended up finding, because there was a control group where
when you typed in GEICO, the ads on the side they had
modified to be Nike shoe ads.
And no-one was confused.
So really what it was, was voted in as being, was that
it seemed to be that it didn't have to do so much with the
keyword as the text of the ads themselves.
That the text of the ads themselves were the thing
determining whether someone was confused, not the fact that
someone had typed in GEICO and might see some other
kind of product there.
And so the court actually ended up holding that use of a
trademark as a keyword, without any additional evidence of
confusion or anything else going on, is not likely
to cause confusion. Which means it's not likely
to be trademark infringement. So, that was, that so far, is
the sort of biggest US ruling we've had on this issue.
And the bottom line is that the court went on to say that
advertisers can use keywords like GEICO, and Google
can have them in there.
The only thing they can't do, though, is, they can't
use them in the text or the header of the ads.
So the creative that Hal was talking about, the court said
you couldn't use the phrase GEICO, the word GEICO,
in ad ad text itself.
Because that, is in the court's opinion, likely
to confuse people.
But you can use it as a keyword to trigger your own ad that can
just talk about insurance or cars or anything else.
And that was, kind of, splitting the baby,
I think, in a way. And the case then settled.
And then that was a trial that happened.
And usually these things go up to appeal.
But it didn't, so we don't know what happens after that.
So the second keyword case I want to talk about is
the American Blind case. And this one's kind of
interesting because there's a company called American
Blinds and Wallpaper.
And they were upset because, again, competitors were
queuing their ads to phrases like American Blind.
But, of course, American Blinds also stands for
blind people, right?
It was actually sort of interesting, American
Blind Skiing foundation, which is not a joke.
It really is a foundation that teaches blind
people how to ski. It's actually really cool.
But it shows there a lot of different meanings to
a lot of different words. So, when people are searching.
When they type in American Blind, you actually
don't know exactly what they're looking for.
Maybe they're looking for things for interior decorating.
Maybe they're blind or they know people who are blind and
they're looking for resources.
In America.
And so it kind of calls into question our presumptions and
our assumptions about what people are searching for.
And the law is trying to sort that out.
And, I don't know if any of these are competitors on the
side, but again, for the sake of argument, the lawsuit that
there would be competitors who would buy of those kinds of
keywords and then advertise in competition with the American
Blind and Wallpaper company.
Now, this is here, in the northern district
of California, it's down in San Jose.

There's no rulings yet. And I just wanted to point
out though, because it's a different kind of keyword case.
Where it's not a GEICO word, there's no English word.
That GEICO means.
But here there are English words that can be plucked out
of the dictionary, potentially, by the trademark owner.
And that's one of the things the EFF is concerned with.
Is that trademark law should never extend so far that people
can't actually use words with technology to describe things
that are just merely language.
Like, if I wanted to write about blind people in America,
I should always be able to take out keywords.
And so, there's a competitive issue between companies, but
there's also a kind of free speech issue, freedom of
information issue, there as well.

OK. Three quick initial notes and then we're going
to jump to copyright.
A lot of people hate spyware and pop-ups and adware, but
there's a lot of similarity in the fights going
on around that. There was a case where 1-800
Contacts sued WhenU, which installs a program when you
type in things in your browser window, it'll pop up an ad of,
supposedly, from a competitor. So if you type in 1-800
Contacts, something like a Lenscrafter ad, or some
other ad would pop up.
And a lot of those lawsuits are going to have similar
legal issues to the keyword searching.
Because, again, it's a question of when and how
can you use keywords. And, actually, in that
lawsuit 1-800 Contacts lost.
Because, and this is another interesting point, is that use
of the trademark was actually going on behind the scenes.
WhenU's program never showed you the trademark.
It just looked at what you typed in then check
its own database and displayed an ad, right?
So it was kind of doing it behind the scenes.
And the court said, well, trademarks have actually
traditionally been things associated with products
that you see in the store. So, usually there's a
brand name on the product. And if this is happening
underneath, behind the scenes, the computer just computing
it, it's not really a trademark issue.
So that was kind of an interesting spin on that.
And then, the other two things I just wanted to mention, as
Hal mentioned, is that other countries are coming to
different decisions.
Especially around this opt-in versus opt-out.
I mean, the way that Google tends to, and Yahoo, tend to
new approach the Web, which is the way that most successful
web companies have, is that it's kind of an opt-out system.
They try and spider everything. They try and capture
everything, they're trying to do everything.
And then if people complain they'll decide whether
they're going to stop doing that or not.
To try and get permission from hundreds of millions of web
page owners is just impossible.
It doesn't scale. You can't do it no matter how
many billions you're worth.
And so the idea is that we'll do it, and if people complain,
we'll start to take it out. And that's getting challenged
in a couple of these cases. And France is definitely one
where now they say you can't use anyone's trademark as a
keyword unless they specifically give
you permission. You have to opt-in for it.

So it's sort of breaking this end-to-end idea
on the internet.
So let's talk a little bit about copyright and then we'll
leave some time at the end.

So there's a lot of things going on in copyright.
I'm only going to focus on a couple of them.
Because they're sort of hot topics right now.
But as you know, a search sends out spiders.
They capture all these web pages.
They cache them, they index them, they provide links.
There's a lot going on now with image search
and image indexing. And and then there's the
Google and Yahoo book projects.

So, just a little quick background on copyright theory.
Since you're all fairly familiar I won't go
over this too much. But the general idea behind
copyright is that it's meant to promote the
creation of new works.
That it's sort of an engine that's supposed to help people
write new things, take new photos make new movies.

In order to do that, the idea is that it's given strong
controls to the owners, the copyright owners, to control
who can do the copying of their works after they create them.
If I make a book who can publish it, who can make
copies of it, things like that.
But on the other side, copyright law has always
balanced that with the idea that the public should have
access to these works.
And to the information within them.
That it should be something of a learning process that we get
from it, so we can enjoy them and we can read them, and we
can use them in various ways.

So, copyright owners get control over most of the
copying of their work.
And these are the exact rights that are in the copyright act.
I'm not going to spend too much time on this because
I want to get through. But the main one issue
here is reproduction. And the idea that they get
to control who gets to make copies and multiply their work.
If it's a book, who can make a thousand copies of it.
In general.

But the problem is is that before the Web, the number
of people who could actually make massive copies
was pretty small.
You needed sophisticated machines, or you
needed a business. The average Joe or Jane
didn't really make copies, we bought them.

But the problem is, is that when you go online, almost
everything you do is a copy, right?
So the internet is basically a huge network, distributed.
And when you send something back and forth across
the internet, you send it to one computer.
Which makes a copy, that sends it to another computer.
Which makes a copy, and so on and so on.
Until it gets to its end.
So, all along the way it's, every time you do something,
you're making copies. So here's just a
home DVR system.
We're trying to track DVDs and camera inputs and whatever.
And every single transfer is making a copy onto some
hard drive or into some memory of a computer.

Backup systems that copy onto mirroring hard drives
make copies all the time.
This is how your email works, if you didn't know.
Is that mail gets copied to a mail transfer agent, to
the next place, the next place, the next place, and
finally back to the user.
So copies are made all the time of everything you create.
Your email is a copyrighted work, and it's getting
copied multiple times.
This is an instant messaging architecture.
Each one of these computers. All the servers, all the
clients, they're all making copies of the messages
they're sending and receiving.

What that means is that every computer device
is a copy machine.
And to the extent that that's true, and copyright law
controls copying of items, it means that everyone is
potentially illegal.
So everyone is potentially in infringement of copyright.
And this is the problem with the internet, is
that it's totally broken copyright, essentially.
Because this old idea we had, well, you needed a printing
press or a copy machine of mammoth proportions.
Now, each little device you own potentially can infringe
hundreds of copyrights.
So, the bottom line here is that in order to legally do
something, copyright law sort of has two outs for you.
Mostly. One is, you get permission.
And the second one is that you qualify what's called fair use.

And so, search engines, who are copying, indexing, and
distributing information to millions of people, making
copies all the time and stuff, have to deal with this.
And they sort of have three ways.
One is they try and assume implied permission to do
all the stuff they're doing. And the give an opt-out option.
They rely on fair use, or they just kind of have blind hope.
Like, oh please don't sue me, please don't sue me.

I'm just going to talk about the first two because the third
one you can figure out for yourself how good that's
going to serve you.
JASON SCHULTZ: Yeah, exactly. American Blind hope. Exactly.

It is as true form of Americana.
So, the idea and the theory behind implied permission, and
this started from web pages is that, well you made
available on the Web, right?
And you kind of knew we were indexing.
And so, therefore there's this implied understanding that you
as a copyright owner of a web page have given us permission.
And what we'll do is, we'll set it up you can opt out.
So that if we receive a notification from you, an email
or something that you don't want your web page in Google,
you're crazy but fine, we'll take it out.
And that's where a lot of these kinds of search architectures
have gotten away with it. And it's worked fairly well.
There have been a couple of disputes, but most people
have agreed to it. The other thing is, any of you
know about robot text headers, robot exclusion headers, it's a
little bit of code you can put in the top of your web page.
And most search engines and other kinds of collectors
will respect that. And not capture the page.
And not include it in their index.
So, they're given you both a human way to do it
by notification, and a technical way to do it by
including a certain tag.
And, so generally this system had worked for the spidering
and the caching, and the indexing and even the
linking to web pages.
Because, quite honestly, the most efficient way to get
it out is just to call them or notify them and
they'll take it out. Or a robot dot text
header up there. If you want to hire a
lawyer for $400 an hour to sue them, you can try.
But if the result you really want is to be out of the
system, most search engines and other providers are going
take you out immediately and then you're done.
And that's kind of a good resolution.
There's also a legal mechanism for this the Digital Millennium
Copyright Act safe harbors, that I don't time to
talk about today. But there's also, so there's a
legal mechanism for takedowns if you give notices.

Let's go into the fair use, though, because that's
where the real interesting stuff is, I think.
Right now, the real controversial stuff.
So, fair use is a very complicated doctrine
in copyright law. And I'm not going to try
and explain it right now in much detail.
But I'm going to give a couple of examples to
give you a flavor for it.
There are all these factors and all these tests and
legal arguments and things.
It's what we do a lot of. And so if you're really
interested, you can check out more on our web pages.
But generally fair use has fallen into three categories.
One is personal non-commercial use, the most famous of which
is time shifting in television.
There's a big case that went up to the Supreme
Court about VCRs.

People were recording TV shows and the motion picture studios
and TV studios sued, saying, that's infringing
our copyrights. We did not give them permission
to copy our television shows and watch them at a later,
more convenient, date. And this company, Sony, is
selling an evil Betamax device that allows them to do this,
and you should shut them down.
This is a report that you will hear.
Now, Google is an evil company allowing people to make copies
without our permission. Everyone's an evil
company, right?
Well, the Supreme Court said, no that personal,
non-commercial use, time shifting to watch at a
later date was a fair use.
For a lot of reasons, like, it feels like
it's the right thing. It doesn't really
harm your market. In fact, it might enhance
it at some point. People might get more
interested, and in fact that's what happened.
The VCR became a huge avenue for TV and motion picture
to be watched on. Now DVDs.

Secondary area which some of you are familiar with
is education and First Amendment activity.
Sort of parody, criticism, teaching.
A lot of the imagines you saw here I took off the Web.
So I'm relying on fair use for that.
Because this is an educational context.
You hear that, Jack Valenti?

But generally that's been recognized that in
classrooms and things, that you can do this.
Especially if it's not commercial.
And then the third, and this is really the sort of more tenuous
area of fair use, which is more interesting, is
transformative uses.
And this has really been, this is sort of part of the First
Amendment rhetoric too.
But if you take something and you make more.
You sort of use it in a way that's socially beneficial,
a way of communication, a way of kind of remixing it to
make it more creative.
A way of transforming its function.
So, for instance, there are a bunch of cases around where
video games are taken from one platform and moved to the
other without the permission of the people who made
the video games.
But so they kind of sued them because they had to make a
bunch of copies to do that. But the court said, no, look
they're not competing with you.
They're not trying to sell video games, they're trying
to make your video games more useful to users.
In a different context. And so that's OK.
And that's fair use.

One of the other things, which is thumbnails, which is
interesting, is that that has already been litigated to
some extent in the courts. As to whether making
a thumbnail image for search is a fair use.
And one court has found that it is.
Although there's some interesting arguments arising
in a new case against Google, about why it might not be.
Well, basically, the bottom line when you look at fair use
is, are you reducing sales of the original work, are you
putting it to some other use?
Are you doing something personal with it, something
which is a free speech thing.
Are you making some sort of new and creative and socially
beneficial use that we like.

So let's talk for a bit about image search.
I typed in wolves, I got wolves.
This is what Google does, right?
Google doesn't ask permission for any of this.
They go out, they spider images.
They tag it with some kind of association with keywords,
and then they display them.

So there are a lot of copyright issues that come with this.
And one is that when you capture the image
you're making a copy. When you turn that image into a
thumbnail you're making a copy. When you store the thumbnail on
a server you're making a copy.
When you display it to people when they type in the keyword
you're displaying a copy.
And then you're also providing a link to the original
page which is sort of helping find a copy.
This is why I'm saying copyright is broken because
doing anything on the internet you run into
all these copy problems.
So, a company Perfect 10, has sued Google.
Perfect 10 is a natural porn company.
All their models are not supposed to be
enhanced in some way.
And they put up a lot of photos behind a firewall where you
have to pay some amount of money to look at them.
And they're complaining because they say, they say pirates are
stealing all these images.
They're hacking into the site, they're coming
under false pretenses. And putting them up on
their own porn sites. They're taking these images
and putting them over on these competing sites.
Perfect 10 is frustrated because they've tried to
sue a bunch of these sites. But they keep going down
and up in sort of, like, wack-a-mole, right.
It's like you can't get them off before they've sprung
up in some Estonian server or somewhere.
So they've gone to Google, because they can find
Google and because Google has a lot of money.
And they've complained. They say, hey, Google, you're
spidering all these sites.
And we give you notice but it's not enough because they
keep popping up everywhere.
We want you to affirmatively filter out these images.
We want it to be where they're out unless we opt in.

They also make really interesting argument,
which I had to mention. Which is that, they say that,
thumbnails that you show are actually substitutes because
we sell little porn images for cellphones now.
And I'm convinced that, in my own opinion that they did
this just for the lawsuit.
That they started selling them for cellphones just
so they could claim this. But they're claiming
it's not a useful thing.
That's not as good as the original.
It's actually is the original because we sell
them for cellphones.
So Google comes back and says, well, wait a minute.
This is an automated process.
There are millions of web pages.
We spider everything. We can't tell what's infringing
or not unless you tell us specifically with a notice.
We don't have time or money or any of this to kind
of figure this all out. And not to mention, it's a fair
use to make an image directory.
That's been decided, and it's a socially useful thing.
Image search helps people, right?
So, go to the bad guys.
The people who are actually setting up the websites, not
the intermediary, us, Google.

So that case is still pending.
No decision. But I will say there's
a hearing on this issue on November 17th.
So, before the end of the semester, we might actually get
a decision about whether what Google's doing is going to be
ruled preliminarily legal or illegal.

So now I'm going to talk about Google print, and then
hopefully we'll have a couple of minutes for questions.

If any of you saw, Google has this Google print thing.
And what they're doing is they're going to libraries
and the libraries are loaning them books.
And they're scanning them in. But what they're doing is
not making them available as books online.
But using the scans as indexes.
So that when you type in a keyword, like gat, it searches
through all the books. And says, oh, here are the most
relevant based on that keyword.
And here are a couple of snippets.
And is for works under copyright.
Works that have a fallen into the public domain
after the copyright term. You do get to see the
whole book, but that's completely legal.
But for this, for books under copyright, it only shows you
little snippets around the word, to help you decide,
that is, according to Google, whether you should
try and buy the book. So you can go to Froogle,
or you can go to out of print books.
You can look at local libraries if you want to borrow them.
There are lots of ways you can actually go and purchase the
book or acquire it legally.
And Google is sort of framing this as an information
helpful tool.

Well, so the Authors' Guild has now filed a class action suit
against Google over this.
And complained about it.
And the big complaint, again, is opt-in versus opt-out.
They wanted opt-in, and Google has decided to do an opt-out
structure, because that's what Google does with everything.
They want to scale to a world-size.
They say we're going do it, and if you complain,
we'll take you out. And Google has been doing that.
If authors complain now, they will take out any
books that they want.
But they won't do it preemptively.
So the guild has kind of made this really silly
argument so far. And we'll see if it gets
more sophisticated. But quite honestly, they're
basically saying, look we make money by selling books.
You actually went nefariously to these libraries
and borrowed them. You didn't buy them.
You actually borrowed them. But Google, you're rich
and you could have bought all the books.
And we would have made more money.
And you make money. We want money, pay us.
That's really basically the argument.
And the way they framed in the law is that, in order to
make your index you had to make a copy of each book.
And stored that copy on your server, and those are copies.
And copyright law forbids copies.
And therefore we're suing.
And this is why I'm saying. Copyright law is broken
in a lot of ways because Google's not selling books.
Google's not, you can't, it's not like file sharing or any of
these other controversies where you could actually get a
full copy of the book.
You get little snippets. And anybody who has the time to
go through and piece together an entire book from eight-line
snippets, I mean, how many people in the world are going
to actually do that, right?
So Google's response is pretty predictably, look, we had
to copy these books to make the index.
And that's the last one I kind of want to end on
here, is that a lot of this comes around necessity.
And whether you believe or not, it comes down to a technical
question to some degree. But did they need to
actually make full copies of all these books?
They say they did. Right, that they couldn't
actually make an efficient index work unless they had all
the texts of all the books.
Because they wouldn't know how relevant things were, where
to search, any of that.
Nobody sees more than a few lines.
We link to where you can buy or borrow them.
And that this is an important public purpose.
And this will help you sell books.
So, there are two quotes.
This is from the Authors' Guild guy, who's basically showing
where the authors are coming from, which is, we
own everything. You don't have any decisions
to make about our stuff.
It's our stuff, don't touch it.
And then here's Tim O'Reilly, who's actually on Google's
advisory board, talking about sort of a different approach.
Which is that actually, it's not that people are going to
steal these books from you. But that your actual
big problem's obscurity.
And that you lose more sales through no-one being able to
find your book, or people forgetting about your book,
than you ever did from people stealing it or copying it.
And that maybe this will actually be a
good thing in the end.

So this is the last slide.
And this is where I wanted to end.
And I don't have time to talk about it.
But when you're thinking about all these policies, what to do
about keywords and what to do about copyright, these
are some of the important things to think about.
We want to maintain an incentive for creativity, but
we also want freedom for people to innovate technologically.
So, the stronger that these IP laws are, the more constrained
innovators can be, which can hurt our economy in
a different way. We want to help the
information markets.
We want some amount of user control.
And that we should think about sort of the public domain and
public access to information.
That's it, thank you.

JASON SCHULTZ: There's a plug for EFF.
AUDIENCE: I have a question about American Blinds and
the GEICO case you mentioned.
So, could you clarify a little bit about why American Blinds
case is not strictly weaker than the GEICO case.
Because in GEICO's case, the court ruled that Google had
a right to sell the trademark GEICO to somebody
who bids for it. The American Blinds case
should be even weaker for that.
JASON SCHULTZ: Right, yeah. So the the quick answer is
that different courts are in different jurisdictions.
And they might agree with each other, they might not.
And so that's what we're going to see.
We're going to see whether the court in California agrees
with the court in Virginia.
And if they don't, it might rise up to a level where
eventually the Supreme Court might even get involved.

AUDIENCE: I have a questions on the Google colors.
Google has the distinctive colors.
Can they be trademark?
JASON SCHULTZ: Yeah. Colors, sounds.
All kinds of things could be trademarked.
The key is whether they are actually distinct
to identifying a particular product.
So, yes, if Google's colors are distinct enough, they can be.
But if they're more general, then they won't be.
And I won't go into the legal doctrine, but there's
a fine line to be drawn. There's a big case about
who owns Corning Fiberglass insulation, which
was always pink. And whether they can trademark
pink for fiberglass insulation.
Because it depends on what consumers think.
Do they associate it with that or not.
AUDIENCE: Hi. I had a question about ringtones.
And derivation of work. So, if there's any
legal precedent. So that if, let's say
I have purchased a CD.
And then create a ringtone from that, and then use
that ringtone on my phone. Is there any precedent whether
that's legal or illegal?

JASON SCHULTZ: Yes, and it's very complicated.
Here's the thing. Is that the digital world has
brought up a lot more, many more markets, for work.
So if I write a song and record it, then, yeah I can
make it into a ringtone. I can do a million
different things with it.
And so copyright law is still sort of struggling with, well,
do we only protect you for the original thing you wrote,
which is the song. Or do we protect you for
every single possible sub-market that you could
eventually license to.
And the courts are kind of split on that.
No-one's decided ringtones directly, so I think it would
quite honestly depend on how much you transformed it.
If it was just a copy of the song or the core of the
song, like the bridge or something, probably not.
But if you remixed it and did a bunch of really interesting
stuff, then right now the courts would probably see
that as a possible way to be OK about it.
It also depends if you're commercial or not.
If you do it for you and your friends, it's different
than if you sell for $2.99.
JASON SCHULTZ: Well, right.
I should say that this is primarily based on US law.
And the international issues get incredibly complex. Yeah.

AUDIENCE: Thanks. So, not to get hung up on
specific cases, but it seems like the lines are so fine and
so varied that the average person probably is completely
out of their league if they're trying to figure out if the
particular thing they're doing is compliant.
There was a story last week in the paper about someone who
shrunk the BART subway map to put it on an iPod, you
probably read about it. And then BART made all kinds
of threats and eventually they took it down.
But it occurs to me, why wouldn't BART sue Google?
Because I'm sure there's something I can type that
will show me an image, thumbnails, of the BART map.
And in both cases it's transformative.
I'm trying to use it for a different purpose where
the original wouldn't work.
JASON SCHULTZ: Yeah. So the difference there is,
so there are two things that are really important
about what you asked. One is that this environment
for innovation that I was talking about, one of the
reasons Google can take a lot of risk is because
it has a lot of money. So, one of the things that
we're concerned about it at EFF is, what about the people who
don't have all that money. So that's something I
definitely think about. Because I think
small innovators get chilled very quickly.
So the person who did the subway maps, I think
they just were annoyed. But they didn't want to fight.
Because they probably didn't want to pay lawyers.
But the second thing is, is that it makes a real difference
whether you create it versus whether you provide
access to it. So, Google doesn't actually
create any of these images, as far as I know.
They're just finding them, spidering them, storing
them and linking to them. So it gives them a little
bit more of a cushion.
Because they can say, we're not actually responsible.
We just found it. And when someone complains,
we'll take down the link. And so that's kind of how
Google, Google might get sued, but they be able to defend on
a bunch of grounds such as, we didn't know it was without
permission, we assumed it was permission.
And once you told us, we took it down.
Whereas the creator of the map, it was a lot more intentional.

That's it? It's all totally clear, huh?
AUDIENCE: I'll add a question.
So, you said that one of the purposes of copyright is to
provide the incentive for creation.
How well-tested is that?
I mean, how well-established is that rule?

JASON SCHULTZ: As an empirical matter you'll have to
ask Hal, actually. Because as far as I know there
have been no real economic studies that have actually
shown that copyright is serving its purpose as
currently structured. But this is the thing,
is that there's a lot of rhetoric around that.
About whether a copyright is in fact.
Because, this is the thing is, copyright came from an era, I
mean, many different eras, when it took a lot of money
to produce many works.
So, to actually create the books and the films and music.
And as we're seeing, digital technology is lowering
those costs dramatically. For some people, not everybody.
So that the equation is getting really murky
from my perspective.
And that we may not need as many incentives any more
to create certain things.
Whereas we may still need great incentives to create others.
And I think the law is really bad at drawing a
very circumscribed map.
It likes to have blanket rules.
And so that's where you seeing a lot of the
conflicts has come about.
But nobody's never really tested how much incentive
I think you need to create.
So we don't know.