Tuesday, May 03, 2016

Copyright Office 512 Roundtable: Voluntary Measures and Industry Agreements

Official description: Voluntary alternatives to and modifications of statutory notice-and-takedown process; best practices; collaborative efforts of content owners, service providers and others to address online infringement, including availability of programs to smaller service providers and creators; cooperation in identifying infringers; graduated response programs to address infringement; efforts to delist or downgrade infringing materials within online search results; participation of third-party providers, such as advertisers and payment processors, in voluntary arrangements; overall effectiveness of voluntary arrangements; educational outreach; government role in encouraging private solutions; and other pertinent issues.
 
Kimberly Isbell: Let’s start with voluntary measures that are helpful.  If so, what are their characteristics that could be replicated?
 
Jonathan Band, Library Copyright Alliance: Voluntary measures taken by a payment processor: for a long time.  Victoria Espinel asked them to get together for standards/best practices.  Cooperated w/int’l anticounterfeiting coalition; working well.  The most significant feature: it was developed by payment processors.  Came together w/best practices; responsive to their needs but also to reach consensus w/in the industry instead of trying to work across industries.
 
KI: how important was IPEC’s involvement? Necessary or not?
 
Band: Payment processors were all doing this already; it’s a highly concentrated industry and there’s competitive pressure. Helpful to have IPEC in the final steps but lots was b/c they were already doing it.
 
Troy Dow, Disney: Voluntary measures are a bright spot.  A number of them, in our comments w/MPAA. Principles for UGC—has worked for setting standards. What allowed success was that we had collaborative sessions; multilateral b/t creators, platform providers.  Started by putting aside difficult legal questions about what the law required and set a goal of UGC environment promoting legitimate creation but also prevented infringement.  Then we tackled some problems.  Included tech solutions; included understanding about © owners’ behavior.  Included an agreement that this wasn’t just a one time set of principles but an ongoing relationship.
 
KI: what circumstances encouraged the players to get together?
 
Dow: Underlying framework of DMCA; everyone was unsure what the law would say. Litigation was a potential route. Legislation was a potential route. Prospect of losing a lawsuit on both sides brought the parties together. 
 
KTC: does the legal framework remain sufficient to encourage continued development of voluntary agreements?
 
Dow: many issues in yesterday’s panels have a lot to do w/that. Narrow construction of DMCA shifts balance away from shared responsibility to rightsholders; that does push away from environment of cooperation.
 
Michael Petricone, Consumer Technology Association: Legit services making it easy for consumers to do the right thing—piracy dropped. British record industry: overall use of pirate sites had dropped a bunch; UK predicts will continue. Spotify has been shown to reduce piracy where it opens. 
 
KI: yet the content owners are telling us it’s still a problem. Possibility: (1) not enough legit services; (2) legit services aren’t enough; (3) we will never eradicate piracy.
 
Petricone: eradicating piracy online is impossible; the key is to reduce that as far as you can. Voluntary measures and presenting users w/wide variety of legit and appropriately priced content.  There are real numbers here and they show a great success in decreasing piracy and increasing access to content.
 
KTC: do you think that on the content side, content owners are focusing their approach on developing legit content, or is that still a distribution challenge online?
 
Petricone: 1998, there was a period of transition, to be expected for any new tech; increasingly embracing the internet.
 
Casey Rae, Future of Music Coalition: 512(i) encourages this. 
 
KTC: why wasn’t it effective?
 
Rae: it really wasn’t tried. Rightsholders assumed they’d pursue their rights as they had done previously.  Grokster: record labels were interested in achieving favorable legal precedent. ISPs faced legit difficulty in identifying works.  Things change.  Ongoing relationship: Credit card best practices; ad exchange best practices; separate copyright notice system.  All came from different situations; gov’t has role to create environment, but doesn’t need to legislate anything as long as info is presented from rightsholders to ISPs. But we need to know accessibility and affordability of tech for small & medium enterprises.  Need to continue to evaluate developing tech.
 
Maria Schneider, Musician: a young musician has 45 million plays on Spotify; never gotten a check more than $60.  It is not working.  Troy’s solution: Disney found a way to come to the table, but for musicians individually there are no solutions. Content ID isn’t available to me.
 
Jennifer Pariser, Motion Picture Association of America: Endorses Dow’s optimism about voluntary agreements as partial solutions.  All of them are flawed in that they only have some players and they can only be somewhat effective. More successful = players have incentive to come to table. They face liability if they don’t. Copyright Alert system = ISPs enjoy immunity if they cooperate w/us in a piracy solution, whereas other solutions, like domain name registries, have been more difficult to work w/ b/c they don’t face liability.
 
KI: we’ve heard that the trend in the course is to interpret safe harbors more broadly. Have you seen an effect on the prevalence of voluntary initiatives?
 
Pariser: hard to say there’s a one to one correlation, but for sure when great cases (for us) come out, ISPs have more enthusiasm for voluntary initiatives. Hopefully BMG v. Cox will help the Copyright Alert system, whereas limited liability for payment processors pulled in the other direction.  Entities have their own reasons to do things. Ad networks: they don’t want their clients associated with garbage sites. But court decisions finding that an operator has no liability are bad days for voluntary initiatives.  [And what makes you come to the table on good days for you?]
 
JC: Cox is not in Copyright Alert system?
 
Pariser: yes.
 
KTC: what role do you see for gov’t?  Response to Petrichone?
 
Pariser: content industry has done more and more and more to make content available. Windows are closing.  Enormous amounts of content available legally, and yet piracy is huge, b/c people still want something for nothing so we need more than licensing. What gov’t can do: courts are different from © Office; the Office could designate specific things as STMs. Part of the problem w/getting sites to adopt STMs is that there’s no agreement on them.  Sound of one hand clapping. If we say Audible Magic is a great solution and get no buyin, it goes nowhere.
 
JC: Parsing definition of STM: what’s meant by the use of “open, fair and voluntary”?
 
Pariser: something not like Content ID: available to the public, perhaps at a price.
 
JC: licensable tech?
 
Pariser: yes—you can’t be too small or too big, as long as you make the right kind of content.
 
Mary Rasenberger, Authors Guild: Voluntary measures are good if they work; problem w/ones to date is that they don’t work for individual creators.  Part of the problem is that individual creators have been left out of best practices, voluntary measures, industry agreements; don’t have ability to negotiate w/ISPs.  Authors are left with notice and takedown & its shortcomings.  There is growing book piracy. Complaints up 600% in the last 5 years; no affordable service for authors to use. Examples w/Google’s Content Verification—you can’t do it as an individual. Copyright Alert doesn’t work; 6 strikes is too much. Voluntary efforts by advertisers aren’t working either—our authors have Google Alerts set up and get dozens a day; they click and they get ads for sites they’ve just visited. Individual complains to payment processors—your notice just gets lost; they only want to deal w/ trusted senders. Individuals should be part of the negotiation.  STM: the tech exists, and if creators were part of that negotiation, there could be potential for great relief.
 
JC: Scribd?
 
Rasenberger: it works, but it’s not readily available to authors. We’d like to see industry adopt something like BookID on a wide basis, including ISPs. In a way authors could readily avail themselves of. Most creators simply lack the resources to spend on additional tech or to hire services to assist them.
 
JC: why not available to individual authors?
 
Rasenberger: they’d have to be part of the service, which they’re not. [I’m not sure that’s true.]
 
KTC: are you saying the fingerprints that filter is not something that individual authors have participated in?
 
Rasenberger: yes, and then the ISP doesn’t want to filter. They’ll take down in response to notice but not filter in advance.
 
Victoria Sheckler, Recording Industry Association of America: Voluntary initiatives can be helpful, but everyone has to get in the game for it to work. Varying degrees of success. Has to be in backdrop of working legal system.  BPI’s reduction of piracy: different legal regime, used differently, creating significant reason for reduction in piracy not applicable here.
 
KTC: In terms of the difference, pirate sites?
 
Sheckler: our comment suggests court orders had significant impact on piracy.
 
KI: are there particular characteristics shared by successful initiatives?
 
Sheckler: building trust, skin in the game, regular communication.
 
Lui Simpson, Association of American Publishers: Successful for those who can afford to be part of the measures—too expensive, not w/in reach of smaller rightsholder. Should be some push from gov’t to make these measures far more effective.  They become successful b/c there’s interest in engagement.  Pushing parties together about what might work: that’s needed.  Petricone puts onus on rightsholder to solve a problem they didn’t create.
 
RT, OTW: Interested in the claim that “everyone has to get in the game”—but what does that mean?  Big website does not mean big notices.  Our website receives 100 million visits/week and gets fewer notices than there are people from the Copyright Office here.  Wikipedia is orders of magnitude bigger and reports similar numbers, most of them flawed.  We’ve heard a lot about sites that ignore DMCA notices (overseas sites, SciHub): making such sites double plus illegal, since on the facts as stated they already are vulnerable to liability under current law, is not costless; it hurts the rest of us trying to do the right thing.  We have experience with a government mandate to use filtering technology: Sabam v. Scarlet case in Belgium: injunction overturned because Audible Magic didn’t work as promised.
 
Nancy Wolff Digital Media Licensing Association : Tech is there for reverse image search, but there’s no risk of massive litigation b/c licensors are small and can’t afford litigation so they won’t come to the table.  Multiple options for legit licensing of images, but it’s very easy to infringe.  Small claims court might help.  No voluntary measures there b/c no reason to talk.
 
Greg Barnes, DiMA: (1) I share the optimism about voluntary measures; allows different people to come to the table and avoids one size fits all approach that would doom us. (2) Gov’t role: important role in bringing people to table as objective facilitator, but shouldn’t put thumb on scale to achieve a certain outcome.  On licensing: Petricone’s point about ability to have licensed content out there decreasing piracy is hard to deny. So many different studies [Australia, anyone?] show this.  There are still problems in licensing musical performances, mechanical licensing—broken for decades, and Office knows this.  Online video services’ ability to stream video has been hindered based on relationship between studios and DVDs.  Lots of work to do, but industry agreements allow us to talk about this and reduce demand.
 
John Garry, Pearson Education: Experience negotiating—tech for websites that can screen in advance—none of the voluntary aspects deal w/outlaw sites, and they’re a tremendous problem w/ no incentive to use voluntary measures. Effective: website that came to AAP early on and wanted that part of this business model; they wanted a business relationship.  Another experience: negotiating w/a large website that looked DMCA-invulnerable; found a chink in their armor so they negotiated to become a legit business. Every great fortune is founded on a great crime. Nice relationship going forward. Voluntary is great when you can get it, but the outlaws are a real problem.
 
Melvin Gibbs, Content Creators Coalition: Garry’s right.  That period of transition has become a permanent state for us.  We’ve explored voluntary compliance.  The parties are siloed and not speaking. ISPs have been lax in codifying standards for accepting notice. True market failure. We want gov’t to facilitate.
 
Thomas Kennedy, American Society of Media Photographers: Simson & Wolff are right. There are organizations that need to talk with individual creators, and that’s not happening b/c there’s not sufficient incentives. Voluntary measures can’t do that.
 
Kerry Sheehan, Public Knowledge: done right, voluntary measures can ensure protection for speech and allow competition/avoid barriers to entry. But we shouldn’t just talk about this as rights owners and ISPs.  It’s the  world of internet users. These agreements need to be voluntary, not the result of coercion, threats of new gov’t enforcement measures.  Need to be from open process, also public interest voices. We haven’t seen public interest participation and these agreements can be unfair to users and smaller providers. If filtering is being proposed as STM, that’s especially important.  A more traditional open standards body would be more appropriate.
 
KTC: considering the public interest: how do we do that?
 
Sheehan: greater transparency, greater inclusion of groups who speak on behalf of public interest.
 
KI: what are the problems? Unavailability to smaller content owners. Other issues w/existing voluntary measures? Do you see a way to fix or improve those shortcomings and what would that look like?
 
Jonathan Band: Definitely preferable to coerced measures.  HEOA: coercive measure. Wrongly assumed that campus infringement rates were higher than elsewhere but that turned out not to be true.  Legitimate study that tries to understand the notice system and people are criticizing it b/c it’s based on a sample!
 
JC: Is it a bad law? 
 
Band: mandated education is a bad thing. No one’s ever been able to show that infringement causes huge substitution; the amount is subject to debate. I’m not convinced that requiring people to watch an online video is really going to change behavior.  Rather, what changes behavior is the fact that the old world where there were creators distributors and users has become meaningless—every user can be a creator; they become more sensitive to the complexity of © and its boundaries.  When you make your own videos, you understand what you’re creating and what you’re using as building blocks, as all creators do.  The act of creation allows you to recognize what you owe and what you don’t.  Educate users about the rights of others? No, educate them about their own rights, which comes naturally.
 
Terry Hart, Copyright Alliance : Not a concern about any particular measure, but more data is always important. Copyright Alert system: overview of # of notices; very helpful. More recently, we had the PTO best practices in sending notices proceeding.  Written comments suggested it has been effective. Worth looking at how well it’s worked. 
 
Rae: Inclusivity is the key need. Look at earlier agreements later codified, like mechanical royalties, or streaming royalties for music that were blessed by Congress. That’s stood since 2000, though it has lots of failures. And it’s inclusive of all those eligible to receive royalties.
 
KI: on multistakeholder process: we heard some people basically saying there were too many cooks.  Is there a way to balance inclusion with getting so big that it becomes unwieldy and you can’t reach consensus?
 
Rae: target the problem to be solved. If looking at repopulation of infringing links, limit to UGC sites, not search (though search is related to that).  Array of tech vendors so you can understand what they do.  Small artists included is absolutely important, and same w/developer community, who’s theoretically the builders of platforms for us.
 
Schneider: what we don’t do in this country. We don’t allow people to make money through illegal activity largely through initimidation. That’s racketeering. For  me, that’s what YouTube does. With these data lords [nice!] of unimaginable size, represented by att’ys and lobbyists that are siphoning my assets.  All the large studios in NY have closed; you can’t record a large film score in NY any more. All over the world—old men tell me how under Communism they listened to jazz and it gave them hope.  This is a culture of literature, of music, that we want. This isn’t about you protecting a large business making money no matter what. Do we want a culture owned by one company? I don’t. Voluntary measures, best practices like fingerprinting required by every company; standardized takedowns; no required agreement to TOS; checkpoints educationally on upload for photography, for music, for everything framed by the Copyright Office; videos that people have to listen too so they don’t have to watch YT’s copyright basics video. Muppets!  Fair use is jiggling around and you can’t read it.  It’s ridiculous. A ratings system for everyone that does a takedown or counternotice. Forces people to have accountability for takedowns and counternotices.  Everyone should agree to it. It’s common sense to anyone who doesn’t have a hidden agenda.
 
Janice Pilch, Rutgers University Libraries: HEOA, regardless of who pushed it, it is perceived to have improved the P2P situation in universities. May seem onerous and rigid but appears to have had an effect.  How could that be bad? Education is important as a viable approach to changing behaviors.  There’s tremendous confusion on right or wrongness of infringement in the context of viral social media messaging that’s anti musician, anti copyright, anti publisher, anti human. Pushed by the industries who benefit most from infringement, translates directly into cash for them. Users benefit from infringement and they have various motives. Sometimes they’re innocent b/c they don’t know or are confused by social messaging; there’s no standard for national copyright education and people never learn.  Students commonly never have heard of copyright or fair use.  Industry-driven social messaging tells them that infringements is a good thing; contrary to basic social instincts and norms to respect others’ works. Education could be stronger. For the public it doesn’t exist in good forms.  But we need to kill the business model of infringement first.
 
KTC: are there studies about effects of education on user behavior?  Social messaging that’s anti-©?
 
Pilch: HEOA requires universities to review effectiveness of plans to combat unauthorized distribution. I haven’t read the reports, but they exist.  On social media, on the basis of personal experience, I see it on blogs and listservs.  We’ve heard of bullying people who object to their works being used, and of people who agree—you get a string of communication beating someone up for liking © or wanting their work taken down. Can’t cite specifics.
 
Pariser: Not enough voluntary agreements—incentive to come to the table. On the educational piece: you’re hearing two different streams of ideas around education—one is we need it/another that the current info sucks.  Copyright Office could create more educational materials for consumers. Becoming creator is not its own education; in her experience, when you tell a middle school student that her selfie is © they don’t understand the plight of the © industry because they want to give it away for nothing [the horror!] and that doesn’t convey the message we want to convey.
 
Rasenberger: Voluntary measures can’t be the whole solution b/c they don’t address criminal pirate sites, the source of a great deal of book piracy. They move around the web and are mostly situated abroad.  Mandated TPMs through 512(i): it would be important for the process to be mandated; given that the burden is on rightsholders, there’s little incentive for ISPs to come to the table. The gov’t has a role in convening these kinds of standards creation in multiindustry processes that are open, fair, and voluntary.  BookID works only with Scribd; a mandated process could force other service providers to also adopt it.
 
Education could help w/some users. Authors tell us, particularly in genre field, that fans tell them that they only read books for free—they have no shame. Free books are so readily available. Need teeth for penalties, just as with speed limits—you need to give tickets to pirates online.
 
Sheckler: In terms of user interest—users are first and foremost in our minds.  That user wants to interact w/our content and we want to teach them the right way.  PK and CDT were invited to work with us on the CCI initiative.  We work regularly through CCI on education for K-12.  I find it surprising for Band to say it’s coercion to follow the law or petition for a change in a law.
 
Simpson: participation has to be broad and inclusive. Payment processor negotiated: rights holders invited were limited, not inclusive. We do face a problem of overinclusion w/o expertise, so it needs to be a balance. Preconsultation measure allowing those to voice their concerns. Need to compel people to
 
Rebecca Tushnet, Organization for Transformative Works
 
Back to the Q: What are the problems?  Content ID: well known problems with overblocking fair uses and falsely claiming revenues owed to others, recited extensively in comments and also routinely reported to us by our creators, including the internationally recognized artist I mentioned yesterday.
 
Different genres: Scribd’s own website clearly explains the two big problems with BookID: [Scribd’s bookID:
 
“BookID relies upon computer-readable text in digital documents. Content scanned from paper sources may not contain computer-readable text data, making those sources unsuitable for use as references. Similarly, digital documents encoded with optical character recognition (OCR) technology may contain garbled or partial computer- readable text data. This may be true regardless of whether the document is readable to humans. These conditions make it very difficult, if not impossible, to detect matches….  [Note that this means that evasion is trivially easy: all you need to do is insert things that people can’t see but computers can.]
 
False Positives
 
The BookID database may contain reference samples from educational textbooks and other works that contain long excerpts of classic literature, religious texts, legal documents, and government publications that are typically in the public domain. This can occasionally result in the removal of uncopyrighted, authorized, or public domain material from Scribd.
 
… Unfortunately, the volume of reference samples and uploads to Scribd prevent any sort of manual oversight or notification prior to effecting removals.”]
 
Result is: overblocking: quotes from public domain materials or even fair use quotes of another book: first uploader blocks subsequent users of quotes; underblocking, just need to scan using OCR. The change could even be something invisible to the naked eye, such as putting a nonbreakable space ( ) in place of a regular space, or adding random sentences and hiding them with CSS. 
 
[More generally, a simple filter is trivially easy to evade: easy to recode media in new file format and get different hash; algorithms to spot minor edits would be difficult both in terms of programmer time and expertise and computational resources. ]
 
If this check is actually supposed to work (that is, catch these workarounds), you get into the realm of plagiarism detectors, which are a fairly complicated technology. It would take us years to develop our own plagiarism detector, especially given we're part-time volunteers.
 
How to fix?  No perfect fix.  Easy appeal, nonthreatening about piracy, walk people through: In some cases, Wikipedia’s fair use and public domain guidelines for use of images would be good places to start: useful for people who are highly motivated and willing to invest a fair amount of time.
 
Not only is this education stuff really something that its proponents imagine being imposed on the unwashed others, rather than on themselves every time they seek to upload a photo to Facebook or send an email, it’s also yet another mirage. We know people don’t read the terms and conditions. We know they (we) just check the box.  [Copyright, which most people don’t care about, won’t be the topic that changes their minds.  There are only a few effective ways of making disclosures, and you can only do them once in a while or people tune those out too.  Education sounds like having your cake and taking it too, but it’s not that simple.  There are things you can do in particular circumstances once the issue has become salient to people, like Wikipedia editors, but the mandate being described as ideal would not be doing the work; what would be doing the work is the associated filtering mandate.]
 
Nancy Wolff: can’t speak to voluntary measures in our area b/c there aren’t any. Takedown doesn’t work. Harassment as a result of notices. Copyright Office guidance on STMs would be good. Certain creators shouldn’t be excluded.
 
Wayne Josel, ASCAP: we spend time educating our licensee base about what the law is. Easier to get people to recognize obligations to take a license before engaging in bad behavior than to correct bad behavior once it starts—better for us to speak to a guy who’s about to open a restaurant than one who’s been playing music for 3-4 years.  Contra Band, opposite of respect takes place when people create new content—the ease of creating overcomes their sensitivity to others’ rights. [How dare these new creators find out what creating feels like.]  User experience overrides information; people don’t click on the terms of service; services want to make it frictionless so you no longer have to warrant that you own what you upload.  [I agree that it’s a problem, but people ignore these!  You can try all you want!  The reason the services want frictionlessness is in part that while friction does deter uptake, it also doesn’t actually leave the people who sit through the legalese with any greater appreciation of the TOS.  So the benefits don’t justify the costs in most cases.  If friction worked in terms of having people internalize new rules, there’d be more of a reason to use it.]  The law should be required reading. 
 
Dow: the ones that work better are collaborative, not unilateral. 
 
Gibbs: public should be encouraged to think of themselves as creators. Creation is built on other creation, which is why it’s in the Constitution. People do need to be educated about their rights—not just fair use, but you made it and it’s worth something. 
 
Sheehan: we should provide meaningful opportunities for public input and transparency in CCI going forward.
 
KTC: education—what’s wrong w/education?
 
Sheehan: consider differences b/t ISPs, resources, user community. One size fits all will  never work—under and overinclusive, with unexpected consequences. Balanced content in educational programs should respect users’ rights to reuse content in fair and legal ways.

No comments: