Panel IV. Campbell and the Future of Digital Technologies (Moderator, Dean O'Connor)
O’Connor: output of one creative person becomes input for the next. Consider how copyright and fair use work in that system.
Robert Brauneis, George Washington University: How have new technologies influenced fair use jurisprudence? (1) Personal copying and distribution and the personal/commercial distinction. Before Sony, there had never been a single published opinion on fair use in private noncommercial copying. Non-networked tech brought personal copying to the attention of copyright owners, leading to fair use for some forms of personal copying. Few decades later, networked computers challenged personal/commercial distinction.
(2) Computer programming/functional works. Pre-Feist, there was a big branch of fair use jurisprudence devoted to ok uses of factual works; that faded as noncreative facts were excluded. But a new branch arose, dealing w/largely functional works and reverse engineering.
(3) Nontraditional uses. Accessibility to the blind.
(4) Dissolution of the concept of the copy. New tech significantly eroded the role of the copy as the intuitive unit of consumption. Fair use will likely step in to mediate between the intuitive sense of exploitation versus formal reproduction. In non-networked world, users typically acquired only one copy. Now many of us don’t know whether we own a copy of a work or not, since Spotify commingles local and streamed; we don’t know how many copies we own or where they’re stored or when another copy is being made—the cloud obscures, faithful to its name. Correlatively, it doesn’t make sense to treat cloud syncing as we do printing another hardcover book. Reframe from copying to access/exploitation? But it’s more than a little difficult to create a new conceptual scheme and harder to get Congress to act; courts will mediate slippage between rights and practices that seem widespread and harmless, and they can call on fair use to do so.
Sandra Aistars, Copyright Alliance: Worth distinguishing b/t creating new copyrighted work and facilitating new tech. Campbell worked very well for the former, but not so with new distribution models/purposes. The beauty of fair use is that it’s both flexible and case-specific; allows courts to consider all the nuance; guard against new set of reflexively applied presumptions w/negative effects on creativity. (Ideological drift …)
Authors aren’t threatened by the concept of fair use, nor do they want to see it limited. Authors of all types need to facilitate the use of existing works, and celebrate contributions tech has brought to their lives. Tech innovators face similar challenges trying to get product/service launched and authors are sympathetic. Seems to her odd, as June Besek noted, that partial copying by an individual artist is scrutinized more than massive full-on copying of entire works by commercial entities, which is presumed to be transformative b/c different purpose.
Intermediate copies are all lumped together. They can be used in ways not linked to traditional sources of income for copyright owners, or simply to design around existing law to avoid licensing. Aereo is a good example of design-around. Intermediate copies used to facilitate new products/services can have different impacts on different market sectors. Search engine that exposes artists to users is beneficial, but taken too far or applied specifically it might undermine the market or supplant the market for visual images entirely. Should distinguish and not say all intermediate copies are created equal.
Do we need different approaches when a new work is created versus new purpose? No. The four factors serve us well, allowing each use to be considered on its merits. Interesting to hear Beebe’s empirical research indicating that it’s easier to predict outcome than some might have expected. She worries that it’s easier to make a decision based on nuanced analysis of four factors than it will be if we allow the trend to compress everything down into one test, bringing a tendency to be guided by perceptions of morality or social utility, which are far harder to predict.
Yoko Miyashita, Getty Images: over 200,000 photographers, 100 million online and 85 million more in storage. When we talk about copyright, we mean creative, technical, highly skilled works that our photographers may put themselves in harm’s way to get. We moved through technologies than now make anyone a photographer and a publisher. The universal language is imagery. Part of that language is “like” and “share.” For younger generations, using images is essential to speech and personality, not just a “nice to have.” Genie’s out of the bottle. We don’t want to be regarded as the ones who want to break the internet and not taken seriously. Trying to take expression away from the public is a losing game.
We tried the enforcement route and we learned the hard way. (Turns out they sent a letter to Corbis, despite cross-distribution agreement.) Unwinnable whack-a-mole, plus potential long term brand harm. So we needed to evolve instead of becoming irrelevant. Getty Images Embed: 99% of content is available for free, for noncommercial purposes. Social sharing has been added to our images. Facebook, Twitter, Tumblr, Pinterest: adds watermark, link back to images, and information on license. We want user traffic, eyeballs, data: the currency of our marketplace.
Our biggest competitor is Google. Among buyers who regularly license content. 9 out of 10 unauthorized users we speak to point to Google as their source of imagery. Kelly v. Arribasoft & Perfect 10: what’s happened in search over 12 years. We went from thumbnails, which used to have little links to the source site; now shows nothing but images like Getty Image search results. Then it goes to a beautiful large view with functions that have had a huge impact on us. The addition of larger images, and arrow features that allow you to toggle from one image to the next, has had a devastating impact on traffic for us. Not going to get into the framing issue (by which she means whether framing implicates any §106 right). Easy to right click and save on a computer; they are the same size on a mobile device. They significantly reduce clickthrough rates for source sites.
This substitutes for eyeballs for all our paying customers who are licensing these images. We are bearing the cost of hosting those framed images when viewed that way. Is it transformative or substitutional? Look at the market harm. Is this search or a wolf in search’s clothing? (I’m gonna go with search.) All we want is eyeballs, data—it’s negotiable currency. Are these intermediate copies designed around existing law to avoid licensing? That’s a key question for us. (And a weakness of the Aereo reasoning: should we say that Google search behaves as if it is making a copy, so we’ll consider §106 implicated?)
Matt Sag, Loyola University Chicago: The dissolution of the traditional concept of copying. Exchange of value b/t author and consumer was made easier/clean by first sale: money for copy. Then public performance entered the field; broadcast as another source of value, so we get broadcast/cable retransmission rights, crowding around reproduction right but not really changing its significance. Then we get tech led by photocopier, disintermediating copying for businesses and consumers. Then goes digital and networked. Now it’s possible to look at when people stopped saying “the United States are” and said “the United States is” by copying the contents of an entire library: a nonexpressive use. The machine is the only one that sees it all. We’ve never had to think about copying not for expressive use before, but courts have handled it rather well. There’s an interesting question about whether nonexpressive use should be considered transformative, or recognize it as a different form of fair use—latter would be cleaner, but Campbell is still important because it tells us that it’s not the loss of despotic dominion that’s a problem, but rather expressive substitution.
Jule Sigall, Microsoft: He’s not really worried about any of the stuff his family does, like using image search for personal communication/creating new meanings within the family. Think of all the things we use: Outlook, Powerpoint, image search, telephone lines, Photoshop, Skype. Tech becomes part of the speech. These are our users. Where should we go to help answer questions about whether tech should permit these uses, how we should market the tech? As the person who represented Kelly in Kelly v. Arribasoft, I got a painful lesson in trying to answer those questions using the basic structure of the Copyright Act—a set of broad rights designed to cover lots of stuff, with narrow exceptions (w/exception of fair use). Presenting it that way left the court with a lot of disturbing questions, and thus justified the use of fair use. You proceed at your peril as a copyright owner if you present a copyright infringement case w/out considering the kinds of things fair use considers—it’s just not effective advocacy.
It’s good for fair use to answer these questions, b/c it provides the most robust set of factors to answer the question of whether an activity is ok. They can be misinterpreted, but they’re better than the other portions of the Act (compared to Netcom, which invented a volitional conduct requirement, or Cablevision on breadth of performance right). Not obviously complete or clear, but good set of precedents. Fair use is right in the norm of predictability for laws that in-house counsel are asked to opine about. When you look how my kids use the tech, there’s no going back to the old ways of “paraphrase or get a license.” Not for them, not for the courts. That doesn’t mean all fair use claims will succeed. But fair use has plenty of applications now.
O’Connor: New tech always popping up, like livestreaming Periscope. Does it matter any more whether there are multiple copies, or just uses?
Sigall: Software has had the longest history with digital copying. Industry is moving to offer products away from counting copies in the offering to the consumer. Office 365 in the cloud. As a business matter, the ability for the user/customer to not think about copies is now the expectation—counting would be distraction/inconvenience.
Sag: it’s not the case that Google Book search à Napster. But my computer backs up to the cloud. I have no idea how many copies of my authorized purchase I have floating around, and neither should the copyright owner. More ambiguity about where the copy resides = need to be less doctrinaire about “every copy is controlled/has independent value.”
Aistars: doesn’t matter unless you’re in a litigation posture where you have no other way to advance the argument other than claiming that there are multiple copies. It’s a distributional argument: how to divvy the pie up. Consumer electronics were innovating in the 80s and 90s, in symbiotic relationship with the “content” industry to sell expensive devices. The business model is now more about giving the content away and relying on the data to sell stuff. Leads to more interaction for the user, but the creative work is subsidizing the ability to make those uses. Not the same as 80s and 90s where nobody was capturing that value; tech innovators are capturing that value by showing ads/using that data. (I think “nobody” here means “consumer surplus,” if I understand her properly.)
Brauneis: space and time shifting are quality enhancement technologies. Fox v. Dish: automatic ad skipping. Suppose all those technologies are fair use. One way of thinking: it doesn’t matter whether the © owner sells you a high balcony seat if through the use of tech you can make it a front row seat without having to pay more. It’s not obvious that the answer is that there should be no differentiation between qualities or levels of access to work to improve social welfare.
O’Connor: we used to think of vinyl as lasting forever, and then it doesn’t. Now we can, in the cloud. But am I buying a forever copy? Or do I lose access if I don’t pay?
Sag: Peter DiCola has a good paper on this, the option value of music. Cassette: really a one or two year license. (How badly does he treat those?)
Aistars: buying a new device to listen costs you something, whether it’s having to watch ads or lost data. If you are buying one, why not pay for the rest?
O’Connor: I don’t want to buy new devices.
Sag: consider logic: if a blind person has glasses that capture an image of a book and translate it into an audio version I can hear. Do I really need to pay for the book twice? The tech has enabled more increased value, but why should some of that value go to the book’s copyright owner for those digital images?
Aistars: would depend on what the manufacturer is doing in terms of gathering info about what the person’s reading, developing additional lines of business that depends on the interest of the underlying work to the consumer. In that case some share should go to the person who’s creating the work that the individual wants to read or hear. (Wouldn’t that share be “the cost of the copy she bought”?) Many people do want to upgrade over time. If you complain about having to buy in a new format, you should be equally upset about having to buy the new device to play it on, and she doesn’t see that happening in this new space. (That’s because of the classic excludability problem. The device is excludable without law; the sound recording is only excludable if the law forces tech makers to design that way. That’s why they feel and are different.)
Sigall: dual purpose tech raises issues (infringing and noninfringing). But since more and more stuff is more like an app on different platforms, it’s less about design of particular app or device or purpose. Challenge in this space is to say: everyone will use this for everything; that’s the world. Trying to segregate what you should/shouldn’t do is extremely difficult. Fair use is relatively good at this. Subset of copyright owners tends to object; what about the rest of the copyright owners whose works flow through the tech? We have really bad info about the preferences of those owners. How do you as a tech provider segregate these, especially without a formalities system? Some © owners don’t speak with all but the tech is all for everything.
Naeve: some © Act provisions are extremely reactive to tech—cf. §110. Fair use is more open ended. Which is a better promoter of © policy?
Miyashita: you have to be flexible. We can’t anticipate 18 months ahead in new tech.
Aistars: both are promoters. You need flexible fair use to account for unanticipated events or one-off situations. There are situations where it’s been beneficial to have a more specific permitted use and in some cases have that use compensated. Would prefer not to have industry-specific terms in legislation, but in some other place. (Presumably a reformulated Copyright Office.)
Sigall: interesting to see whether a plaintiff owner argues at a high level of abstraction where the tech doesn’t matter versus arguing about which server makes which copy. His sense is that it goes both ways all the time.
Brauneis: the longer a delay there is in legislative action, the more we will use the courts to make new law.
RT: public statistics on uptake of Getty Images? Is Getty sending C&D or non-C&D letters to people who aren’t using it to switch people over?
Miyashita: we are sending letters. Encouraging/highlighting the availability. (But no answer on statistics on uptake, sigh.)
Heald: thought that Kelly is defensible but appalled by current Google image search. Is anyone litigating it? Second, Hargreaves (UK) report—no changes in the absence of empirical data (RT: though note that the initial law was not adopted with the same rigorous standards). Is that a good idea?
Sigall: It’s hard to argue against policy based on empirical evidence, but the question is how you get it in timely and effective fashion. Can be very difficult.
Q: did Getty have photographers object to the Getty embed feature?
Miyashita: no. We have to acknowledge practical reality of the types of uses for which embed and social sharing are ideal. Attribution, traffic, eyeballs, data, linkbacks start to make sense as overarching strategy.
Loren: also have questions about new image search. Are you doing anything? You have market power. You could choose the robot exclusion header. Setting the law aside, there’s a business reality.
Miyashita: we could use robots.txt, but the 100s or 1000s of other licensees—there’s no means for them to exclude; publishers would have to implement robots.txt, and then the stolen sites would be the prominent results. (That’s … not exactly true in that the publishers could implement robots.txt just for images, and allow text to be indexed. And then there is the DMCA.)
Aistars: many people we work with are individual authors. To the extent there are ways to aggregate and get value from works, whether monetary or data/interaction, those are positive things.
Sigall: has seen interesting things happen around accommodating multiplicity of uses, fair and unfair. Videogames. Valve’s Steam platform: starts from ground-up understanding of what customers and users and creators all understand about what you can and can’t do. They accept fair limits because the platform has been built that way. Create norms around usage that are more meaningful, enforceable, and practical than if you looked to the Copyright Act to mediate the transaction.
Samuelson: the role of TPMs and anticircumvention. When I read §1201 submissions from GM and the Auto Alliance, they have concerns about battery life and warranties, not about infringement. In the internet of things, TPMs may become a new set of issues; fair use circumventions already happen every day. How will TPMs and anticircumvention evolve?
Sag: most TPMs don’t work. When they’re circumvented, often won’t give rise to liability, for courts following Chamberlain and Lexmark. Attempts to use TPMs to control aftermarket could be an antitrust violation if you buy a single-brand lock-in theory. Lots of moving parts, and then there’s whether the Copyright Office should be writing regulations. (Um, the answer is clearly no: what business of the Office is it how long the battery of your car lasts?)
Sigall: there’s a role for TPMs to preserve the rights of copyright owners. Valve uses TPMs in a pretty significant way, but they’ve built it to be flexible. Tech is deployed in a social and communal setting that determines whether or not the tech works; the legal architecture sits in the broader community. If you do it right you diminsh the need for circumvention. You will have abuse of tech, abuse of legal protection. Safety valves: the © Office rulemaking, which gives us information about what’s working and what’s not. Chamberlain and Skylink also showed courts willing to create exemptions that didn’t exist in the statute. Don’t paint too broad a brush: TPMs aren’t going to save copyright owners or destroy users. (Just uses.)
Brauneis: in a networked world, when you can keep important parts of the works on a remote server, the use of TPMs is less important.
Aistars: TPMs can reassure authors into supporting new formats. Problematic when used for printer cartridges and garage door openers; courts have been able to get it right.