Thursday, July 23, 2015

ISHTIP at Penn, Part 5

Session 3 | Marianne Dahlén (Uppsala University, Sweden), Moderator
 
Design and Copyright: An Open Question?
Stina Teilmann-Lock (University of Southern Denmark)
Commentator | Jessica Silbey (Northeastern University)
 
Openness in the law for fair followers?  Design in law in Denmark from 1908: fair followers can copy, despite existence of copyright protection—a balance b/t copying to develop and proliferate design and control/remuneration for designers.  Balance is very hard to achieve; pendulum swings between 1908 and 1960 due to categorical shifts in legal reasoning.  Rely on foundational assumptions about art, purposes of ©, and people © is supposed to benefit.
 
One approach: deny © to industrial design of any kind.  Platonic idea of art. Other trend is protect original ornamentation wherever found, but make it thin/identical copying only. Similar trend in US/Bleistein’s anti-discrimination principle.  Danish SCt, 1908, ruled that coffee pot was uncopyrightable—it was meant to proliferate in the market despite its decorations/artistic qualities.  It had a prototype. 
 
Modernist art: fashion and painting, font and furniture, art and politics were all linked. When Demnark amends copyright act in 1908 to cover prototypes for industrial art and handicrafts, whether or not reproduction takes place w/purely artistic purpose or industrial purpose/practical use: a kind of leveling, bringing more artists under the copyright tent. Similar to US: the copy is an individual reaction to something in nature; virtually any personal reaction will be copyrightable. Also comports w/political movements of modernism/industrial capitalism. 
 
Kantian influence over Danish © was strong and resisted this impulse. Talent or genius, used to reinforce hierarchy of fine art.  Not clear whether paper celebrates this exclusion, though calls it a chance for “openness.” Clearly based on an elitism/certain objects are only marginally protected, while jewelry and other ornamental features were protected.
 
Bauhaus chairs from 1930s: experts found them worthy of protection, but Danish courts didn’t—unique features came from style/particular materials and thus not artistic worthy of protection.  Functionalist furniture + rhetoric of modernist movement made applied art even more problematic in © terms—the binary of fine art/craft is based on this fiction of individuality/common practice also being a binary.  Mythology of originality/genius that denies/suppresses evolutionary practices while simultaneously denying to everyday creators the benefit of legal protection.
 
So broadening the categories of protection to add applied art in 1961 didn’t flatten the hierarchy.  What does this mean for industrial designs?  Simple/common forms could be copyrighted when they existed in harmonious unity, selected by the designer.  Where then is the openness? Disappearance of shared community, mutual following. Savior of openness in Denmark isn’t an antidote to romance of individuality, but rather a doctrine of thin copyright.  Low originality requirement must have as its concomitant a narrow scope of protection.  But then does that perpetuate hierarchy of copyright genres?
 
Should always ask who benefits from regimes and where the harm is.  Openness allowed: benefits of IP equality/leveling down extend beyond traditional authorship to audiences and fair followers.  Progress comes through copying; fair followers need as much openness as possible.  Category of design generally is a problem across IP disciplines.  Whether design is protected in © is also imported into TM and patent. It is a special category in almost every statutory IP regime and its specialness is confusing—a sui generis category w/o justification where there is so much overlap.
 
Teilmann-Lock: Double status of design comes from its basis in engineering as well. We mean different things when we say design—technical and artistic. See it too in the Berne Convention, where it’s left to individual nations to deal.  Conceptions of design have been very different across time/place.  Actors w/the most to say in defining design in Denmark have been graduates of Academy of Fine Arts—furniture architects, a very loaded term in Danish b/c its connotation is a particular generation of designers who made Danish Modern design globally known.
 
Ornament is a naughty word in the design world, which is trained that form follows function.  It means bad taste. Illusion that object can be stripped down to its function. 
 
Farley: in terms of history, Danish design’s heyday was post WWII to late 1960s.  Change in law of 1961, if consequential, comes at an interesting time.  Is there a consequence for design?  Danish courts became good at seeing the art in functional design; not all courts did.  American/individualistic approach—we don’t ask, can I see art in that chair? We say instead: what were the alternative designs? Were artistic choices made?  Danish ideal: reduce to its essence; that would take the US choice approach off the table.  Danish court says artistic design is “naturally” motivated where US court would see that as a reason not to grant protection. When the Q is alternatives, there are almost always alternatives; but if ct is forced to assess art, it may deny protection from fear of having to make artistic judgments.
 
A: Court-appointed experts play a role in the Danish cases.  Experts had managed to persuade courts of their understanding of aesthetics.  Almost lecturing the courts for the first half of the century about modernist aesthetics and finally courts took it in.  (Similar thing arguably w/appropriation art and the last 50 years of fair use in the US courts.) 

Danish designers always have their names on products: a French press is called a Bonum (sp?) because of the name of the designer.  A TM too, of course.
 
Challenging The Black Box: On the Accountability of Algorithmic Law Enforcement
Maayan Perel and Niva Elkin-Koren (University of Haifa)
Commentator | Maria Lillà Montagnani (Bocconi University, Italy)
 
Algorithmic © enforcement by online intermediaries. How/are they held accountable for what they do?  Framework against which we can judge them.  Tech has always assisted legal enforcement; not aware of it most of the time.  Speeding cameras.  In traditional enforcement situations, decisionmakers make a decision and the tech just helps implement it.  But online, private entities translate the law itself into an algorithm.  Functions that were actually discrete are now carried out by the same entity: law enforcement & adjudication.  And the algorithm is unknown, black box. We can’t know if the law is actually being complied with, esp. in situations of fair use. Dangerous effects on public sphere.
 
Proposal: public scrutiny, not judicial scrutiny.  (Could in the alternative have an ASCAP-style antitrust control.)  How to distinguish from content management decisions made by intermediaries as part of their business operations?  Does the public have a sufficient opportunity to challenge decisions? Can they correct erroneous decisions? 
 
Proxies for this: public literacy through transparency.  Due process—ability to challenge decision/have a voice.  Public oversight: ability to contest removal/restore content.
 
Regulated algorithmic © enforcement, as distinguished from unregulated.  The DMCA doesn’t achieve these goals.  DMCA: uploader is not informed when link is removed.  Counternotice doesn’t preserve due process b/c the content is immediately removed: an extrajudicial TRO based only on © owner’s allegations. If this is true of statutory © enforcement, it’s much worse with completely unregulated/private/voluntary regimes—e.g., filtering that prevents publication in the first place—no transparency; no notice; no due process; no counternotice.
 
Google: ex post and ex ante measures, filtering as a business model. Enables © owners to monetize other parties’ uploading. But that doesn’t meet the framework for judging accountability.
 
Barriers to oversight: Technical barriers linked to nontransparent algorithmic mechanisms as such.  Legal barriers: bars on reverse engineering, research: anticircumvention laws; requirement that you aver ownership in good faith (so you can’t test the operation of the system); user-generated barriers b/c users tend not to send the counternotices b/c the wording of a notice is so scary.
 
We know that these systems don’t work from an accountability perspective.  Public oversight would be better than nothing, but why not change the approach: if © and tech don’t work together, start from scratch, and think of something that works.  Doesn’t make sense to try to make the tech fit ©; make © that fits tech, like privacy by design.  More collaborative approach b/t regulators and intermediaries.
 
Elkin-Koren: purpose was to map issues surrounding algorithmic enforcement, and offer theoretical framework for thinking about it. Not necessarily providing ready-made solutions. Context of a larger effort on algorithmic enforcement. We have data for 3 years about enforcement by algorithm and enforcement in court—Israel is a small environment that allows a population study.  Software for example operates only by notice and takedown, not in court; other types of works are in court and not by takedown.  Most interesting finding: algorithmic enforcement is 7000-8000 notices compared to 100 lawsuits over the same 6 months. We are also looking at the notices.  50% of notices are actually related to the right to be forgotten, though filed under DMCA.  Google’s Transparency Report doesn’t give this info—requires a lot of analysis.  History can help understand where we are.
 
Perel: Empirical study that spurred this paper: we tried to learn systematically about how online platforms in Israel enforce ©; whether they verify rights claims; whether they correctly remove only infringing content. Tried to upload different materials, some clearly noninfringing and some clearly infringing (like an episode of House), and some fair use (baby dancing in Lenz-style video).  Recorded results of sending takedown notice.  Algorithmic enforcement is chaotic in Israel: most platforms did nothing to verify rights; some took down noninfringing content and some didn’t take down infringing content. Couldn’t do this experiment in the US b/c of the DMCA legal barriers (anticircumvention and oath requirement in takedown; CFAA). But anecdotal evidence of same errors in the US.
 
Elkin-Koren: skeptical both about privacy by design and © by design: doesn’t help us avoid the challenge. Don’t give priority to people who design the wires. The community should decide, using an appropriate decisionmaking process. Legislators would have difficulty designing the system too.  Challenge: how to design legal interventions that would be more appropriate for this dynamic environment. We could set standards, but those are problematic as well—can create distortions.  We are dealing with continuously changing tech.  Ongoing legal intervention is therefore required.  Platform behavior is constantly changing.  Some of our data we’ve shown to Google; they had no clue.
 
RT: I’m going to do the lawyer thing and ask for specific solutions.  So frustrating; we’ve been saying these systems don’t work for years and now, instead of any proposals for improvement, from © owners we get the Frank Luntz-style phrase “notice and staydown” as their new euphemism for filtering.  What if anything is the most effective way of making these concerns persuasive to non-IP scholars?  Multistakeholder in my experience means: we are going to wear you down with procedure and time, b/c you aren’t getting paid by the hour to represent the public interest.  Experience of recent DMCA best practices statement “multistakeholder process” hosted by PTO/NTIA: most anodyne results possible (only anodyne results were possible).  Multistakeholder process only works where there is ground for compromise: what is that ground?
 
Elkin-Koren: Multistakeholder regimes are not what we’re trying to advocate—there are indeed many problems.  © is just a test case for online disputes/algorithms.  Maybe some solutions can be achieved more broadly across regimes.
 
Q: FB algorithms that vote down stories that don’t help FB financially—serve no purpose other than as a mask for unaccountable power/to encourage brands to pay FB for access.  Similar issues.
 
Q: YT’s US terms of use don’t cover reverse engineering of takedown algorithm, but Israeli terms of use specifically say you agree not to interfere w/© security or inspection mechanism. Another type of legal barrier. Difference b/t jurisdictions is interesting.  User-generated barriers: has been approached by musicians that received takedowns after trying to monetize their channels.  They are very scared.  If you counternotify on grounds of fair use; the algorithm shuts down; maybe that’s another way to challenge (this is not universally true! I know people whose fair use counternotifications have simply been ignored).
 
Jaszi: at least two black boxes. One is the algorithm. Another Q is why the counternotice provisions are so dramatically underutilized.  You categorize some possibilities in the paper: fear, risk of exposure, simple lack of information, lack of solidarity (people feel like exposed isolates rather than part of a group), indifference (it was one of many videos; it’s already up on other platforms). There is a practical point of intervention: we need information about why people don’t counternotify; we have wonderful hypotheses but need real empirical work. If enough made use of it would throw system into disarray.
 
Perel: it’s impossible to send 1 million counternotices a day.
 
Jaszi: but why are we sure that resistance would require numerical proportion?  1000 might be enough.
 
Elkin-Koren: there is no one you can deal with if you have a FB problem in Israel.
 
Q: responses to EU consultation on ©: a lot of responses from intermediaries on increasing responsibility for blocking infringing content. Adopted discourse of economic transaction costs—burdensome introduction of new requirements.  Intermediaries are reticent to handle interactions w/public. Why won’t intermediaries reject transparency on grounds of transaction/operating costs?
 
Q: © as test case for other areas of law—surveillance machinery is the same.
 
Elkin-Koren: transparency is insufficient; comes with costs. I wouldn’t require more reports/more information: intermediaries should allow their systems to be more transparent, open to inspection/monitoring by outsiders as we are being monitored by them. Not necessarily the same type of cost.  Maybe the reason © disputes turned into political disputes is the fact that © enforcement infrastructure is the infrastructure for other types of surveillance and control.  When you want to remove a documentary on rape in India, you convince BBC to file a takedown.  That was the only way to remove it.  It’s not b/c © fits but b/c © is perceived as neutral.

No comments:

Post a Comment