Showing posts with label google. Show all posts
Showing posts with label google. Show all posts

Wednesday, July 20, 2022

Google public policy fellowships partnering with different tech policy organizations: applications open

Google has relaunched its Public Policy Fellowship program. This fellowship is designed to support organizations at the forefront of Internet and tech policy and enhance their research, analysis, and advocacy efforts. Google is partnering with the 12 organizations listed below:

  • The Aspen Institute

  • Bipartisan Policy Center (BPC)

  • The Cato Institute

  • The Center for Democracy and Technology

  • National Taxpayers Union Foundation

  • Open Technology Institute

  • Progressive Policy Institute

  • Public Knowledge

  • R Street

  • TechFreedom

  • Tech Policy institute

  • UnidosUS

The application is currently open. Interested students should apply to the fellowship before Friday, August 12th using this link. 

Friday, March 03, 2017

Reading list, good titles edition

Kacy Popyer. Note. CACHE-22: the fine line between information and defamation in Google's autocomplete function. 34 Cardozo Arts &; Ent. L.J. 835-861 (2016)

But why no publicly available copy, Cardozo Arts & Ent. L.J.?

Sunday, May 01, 2016

FESC: Regulation by Internet Intermediaries

Regulation by Internet Intermediaries
Moderator: Jack Balkin
Emma Llanso & Rita Cant     “Internet Referral Units”: Co-Option of Private Content Moderation Systems for Extralegal Government Censorship
Kate Klonick, discussant: Most UGC platforms have content standards to which users agree.  Impermissible nudity, hate speech, violent content, harassment, etc. can be taken down, and this is protected by §230.  Platforms have become good at moderation of TOS-violating content; they’re much better than the gov’t, and also unlike the gov’t they’re allowed to do it.  UK created an internet referral unit in 2010, dedicated to flagging “terrorist” content on social media sites.  IRUs are becoming increasingly popular for countering terrorist propaganda in Europe; US is starting to talk about that. Blur public/private lines.
 
Authors argue that flagging, and subsequent takedown, constitutes gov’t action—takedown is directly causally related to the flagging. They also argue that IRUs raise the specter of state action: allowing private action to be held to const’l standards where state affirmatively encourages actor; §1983 suits would be dismissed b/c of §230 immunity. 
 
UK definition of extremism: vocal disrespect for key values, calls for death of armed forces—would be overbroad/unconst’l in the US. Also lack of procedural safeguards.  Internet cos. generally don’t provide adequate notice/right to appeal.
 
Klonick suggests that Google, FB, and Twitter aren’t very vulnerable to gov’t pressures and easily able to push back.  IRUs couldn’t do their job w/o social media—the platforms then have a lot of leverage in terms of denying gov’t access to ToS.  [Though the platforms also trade off interests and may sacrifice some for leverage on others, like stopping DMCA reform making them more liable.] Also, content moderation at the big three at least offers an extraordinary amount of notice—what more would you want, other than notice of ToS, notice of takedown and reason, notice of existence of appeals process, all of which are provided?  A lot of appeals processes may not be available per platform, but many times there are.
 
Transparency reports are a good idea: gov’t requests by month, which would let the public know about the relationship b/t the platform and the gov’t.
 
Llanso: we’ve seen some high level examples of Google, FB, Yahoo taking affirmative stances against overreach.  Concerned that there’s no guarantee they’ll always be able to take that kind of stand, or that other/smaller intermediaries will be able to make that stand.  Pressure on credit card companies as an example: Backpage v. Sheriff Dart.
 
Molly Land     Human Rights and Private Governance of the Internet
Rita Cant, presenter: Reminder that 1A isn’t universally applicable as sometimes we think in the US.  Even immediate removal of defamatory content upon notification of same in the European Union can lead to liability.  Institutional views drive courts’ understanding of roles of intermediaries. European bodies aren’t dismissing the concept of intermediary protection for speech, but holding them liable when they have capacity to police users.  If they fail to do so, liability attaches.  View that big hosts have the power to remove content and thus the responsibility is just wrong, according to Land.  Rather than liability based on size, human rights law prescribes a different principle: an intermediary that participates in creation of culpable speech is different than one that merely serves as a conduit.  Regulating them as colluders/contributors to human rights violation is not regulating them as intermediaries. This prevents over-takedowns.
 
But is a platform’s takedown of legitimate speech a human rights violation the way that facilitating murder or illegal mining is?  Those seem very different.  Wrongful takedown of expression: is it even a violation of human rights at all if it’s according to a company’s standards?  Co-regulation for allowing gov’t to affirmatively protect those rights may be quite difficult—generally, gov’t actively enforcing human rights online has undermined those rights.
 
Land: Would like a bright line rule, but jurisdictions vary; would not be seen as legitimate in many jurisdictions giving more weight to dignity, protection against discrimination, etc.  Easily convinced that co-regulation can be worst of both worlds, with lack of transparency.
 
Kate Klonick  From Constitution to Click-Worker-The Creation, Policy, and Process of Online Content Moderation
Presenter: Molly Land: Deal with actual empirical evidence about how this works. Interviews executives and click-workers enforcing the policies: FB, Twitter, Google.  May have just 2 seconds to look at each piece of content.  Transborder nature affects these policies too.
 
For future: closer connection between normative questions and empirical research.  Right now there are a lot of possible normative questions you could ask.  Maybe empirical research helps us understand the nature of the problem. Who’s doing this, what the content problems are, etc.  If these policies are in response to user pressure v. gov’t pressure/avoiding regulation, we might have different reactions.  Promoting v. defeating user preferences may differ from regulatory perspective.  Also, why these companies and not others? 
 
Klonick: FB, Twitter, Google are continuously operating for a while. Primarily UGC companies, specifically YouTube.  History matters: created mostly by small group of lawyers who were committed to the 1A and wanted to take down as little as they could while retaining an engaged userbase—some of the same lawyers moved from company to company.
 
Balkin: intermediary liability rules are state action rules.  The only Q is whether the free speech principle you use prohibits what’s being done.  If we tried an IRU in the US, though the ISP is permitted to have a TOS, gov’t is probably not able to say “please enforce your TOS with this class of content.”  Line could be different—could be Grokster-style inducement for everything, not just IP.  Leaves a wide swatch open.  Internet company could decide to be a passive conduit for something, but also curation and hosting. Don’t want to tell a company what kind of business model it can adopt.  Drawing a line like Grokster offers more opportunities for innovation.
 
Whenever the gov’t shapes the innovation space and permissible rules about when a private party we rely on for speech will be held liable, the gov’t is always already involved in that decision. The human rights laws are always invoked; the only question is the substantive one: what do those laws require?
 
Q: Gov’ts across the globe are resorting to self-help w/data localization and content regulation, often affirmatively objecting to US approach. Art. 20 of one treaty outlaws hate speech: advocacy of hatred that incites hostility, discrimination, or violence.  Microsoft’s response: commit to obey local laws where we do business, informed by int’l law.  We have to distinguish social media from search engines.  Mapping all info on Web is critical part of research/advancing knowledge; rely on notice & takedown rather than looking for affirmatively offensive content. Nobody elected us to make these decisions. We couldn’t hire the right people across the globe to make a nuanced decision.  So we use notice & takedown; we publish our standards.
 
Land: if we just went w/users, it’d be all porn, so it makes sense for companies to have freedom to shape their own communities. Signal to gov’t about where they are going too far.
 
Balkin: consider Southern gov’ts cooperating indirectly or directly w/private entities to enforce private segregation—also intermediary issues.
 
Abrams: terrible terrorist attack; gov’t learns that the perpetrator just watched a particularly explosive and incendiary work touting jihad.  President calls in Microsoft & Google etc. and provides list of things they ought to do to screen out bad content, though you don’t have to do it.  For the safety of the country you should do it; and the President tells the public that she has called for this action on their part.  Is that a problem?  [Note that this already happens in less fraught circumstances—consider the gov’t’s organization of the best practices in DMCA notices.] 
 
Q: when the NYT decides not to publish an article b/c the gov’t pleads with it to hold off on national security grounds, is that state action?
 
Balkin: that’s the Backpage case. 
 
Llanso: reminds her of Innocence of Muslims.  Yes, that’s improper for the gov’t to do, even in emergency circumstances. There are options for more formal procedures.  Telling the country about the request starts feeling like coercion.
 
Klonick: Innocence of Muslims was a big deal for her interviewees—often called the cause of the Benghazi attacks; people took it as incitement (even if that wasn’t really the cause). Even w/pressure from White House, they uniformly decided not to take it down.
 
Balkin: the fact that people violate the constitution isn’t an argument, it’s just a fact.
 
Lyrissa Lidsky: where does gov’t speech come into this and the gov’t’s right to express its opinion?  If Obama had an op-ed saying ISPs shouldn’t publish Innocence of Muslims, is that gov’t speech and ok or not ok?
 
Llanso: fact-intensive: expression of opinion versus suggestion of consequences/coercion.  If they start talking about modifying §230 if website owners aren’t more responsible, that might be coercive.  [But if they talk about amending §230 at some time there hasn’t just been an attack, that’s ok?]
 
Klonick: YouTube took down anti-Thai monarchy videos; claim is harm to the Thai people. Exporting 1A standards in many cases, though.
 
Q: 4.5% of world population is covered by the 1A.
 
Klonick: the click workers are from countries that don’t have an easy context for the n-word. So they have to look at a report by looking at the person’s whole page.
 
Llanso: when content is illegal, transparency about it and where the locations are can be difficult, such as child porn, or avoiding providing personal data.
 
RT: Not a hypo about gov’t pressure: 512 hearings, Katherine Oyama of Google gets directly told: do more for copyright owners or we’ll have to change this.  Did that violate the First Amendment?
 
Balkin: Not a threat if they have a right to do it. Congress has the right to change the rules of liability, unless the reason is viewpoint based.  That’s the fact question to be resolved: whether the reason is viewpoint based. 
 
Q: don’t assume private corporations are benign compared to the gov’t.
 
Klonick: real name policies are designed to make sure people know who’s attacking them. This is a way to control libel etc.  But they differ platform to platform.
 
Llanso: that’s a controversial policy; it also generates worse outcomes for people with traditional Native American names whose names aren’t recognized as “real,” or people at risk of stalking, harassment, or abuse.  Still sees gov’t effort to restrict publicly available speech as more dangerous than indiv. co. decisions, which can have big impacts but not as big.

Saturday, April 30, 2016

Freedom of Expression Scholars Conference at Yale: Search Engines and Free Speech

Heather M. Whitney & Robert Mark Simpson: Search Engines and Free Speech Coverage
 
Discussant: Heidi Kitrosser: Are search engine results covered by the 1A? Larger question of how we decide what’s salient is part of the paper.  Survey relevant 1A case law about algorithms & results.  Of the handful of cases to be decided, courts basically have accepted a rather simple analogy between the algorithms that pop out results and the type of editorial judgments at issue in cases like Tornillo and Hurley.  Issue has mainly come up in unevenly matched cases—pro se or poorly resourced litigants against tech giants.
 
There’s a long temporal and intellectual chain between the initial creation of an algorithm and the spitting out of results—a disconnect in a way that makes it not at all clear that the programmers are the “authors” of the output in the way one puts together a parade/editorial page. Some algorithmically mediated content might be analogous—for example, Stuart Benjamin says, a giant digital billboard calculating the national debt; the calculation is algorithmic but the message overall is expression.  It’s just not always the case that algorithmic results are the product of editorial judgment. Demonstrates weakness of analogical reasoning generally in determining 1A coverage. Point is to get at the core Q: what are the values, reasons, normative concerns that lead us to accord special protection to speech (or other things).
 
Speech/conduct distinction doesn’t resolve the problem; go back to free speech theory for coverage of new tech outputs.  Pluralistic/democratic participation values: Consumer protection laws and antitrust laws might appropriately cover algorithms; algorithms themselves may repress speech if they submerge certain perspectives.
 
Reflections: push them to consider more of a defense of democratic participation as a focus, and/or run through other theories and explain how those would work.  Are you applying theory to determine protection, or to determine coverage?  If you use your theory to determine protection, then aren’t you collapsing coverage and protection?
 
Simpson: More difficult to make decisions here than the courts/some of the scholarship would have you believe. There should be no ready conclusions w/r/t coverage.  Algorithms are neither inherently dissimilar or inherently similar to editorial judgments.
 
Whitney: Analogies by themselves aren’t doing the work because you need to figure out what makes the analogy relevant.  Theory can help figure out if we’re ever going to have limits on the deregulatory turn.  That is a view that one can take of the 1A, but then you have given everything 1A protection (or coverage).
 
Jim Weinstein: Courts have intuitive, unarticulated theories they use when they analogize; good to bring them to the surface.  You suggest possible fairness doctrine, but even if something isn’t covered, the justifications themselves can trigger 1A concerns: chocolate isn’t speech, but if the gov’t wanted to regulate chocolate because consuming it made people more likely to oppose the gov’t, that would be a problem. Gov’t intent to avoid an echo chamber online: same Q.   [Note that this position may imply that federal mortgage insurance is a violation of the First Amendment; the justifications at least included that being a homeowner increases people’s involvement in the community, self-regard, and motivation to work—cf. the more recent discussions of the “ownership society.”]
 
Deven Desai: if the NYT started to use algorithms to take on Google News—would it lose protection?  Facebook, Yelp—is there a search engine difference or not?  TripAdvisor and Yelp v. Google—you can’t get around Red Lion that easily. If there’s not a scarcity, then people can choose something else.
 
Simpson: Algorithms on their own don’t determine anything about coverage.  Making a claim about how courts shouldn’t be thinking about algorithms, not about how they should be.
 
Enrique Armijo: When you accuse Google of inconsistency on net neutrality. The ISP argument is that they want to reserve the right to edit; Google’s position is that they edit all the time, so he thinks there’s a fair difference.  Regulating GM more easily than Google: but what about Target?  When I go to Google, I’m looking for speech, but when I go to Target I’m looking for avocados.  (Hmm. Many times when I go to Google I’m looking for, well, avocado-colored suits at least.)
 
Jim Tsetsis: What about the Press Clause?  If we treated it as having a separate meaning, as the SCt has not, then we wouldn’t have to sweat so hard about the difference b/t GM and the NYT, and could think better about Google. 
 
Whitney: Hurley line is an issue; not traditional Press. Would require quite an intervention from the SCt; would Hurley come out differently even if we separated out the press?  And the search engines would be fine saying that they were like the press, and only use editorial judgment argument as a backup. The issue is still analogical: what makes something a “press.”
 
Robert Corn-Revere: it’s the organization of information, so what about that isn’t protected by the First Amendment?
 
Whitney: The outcome might be coverage/protection; there are coherent accounts that would include search engines, but also coherent accounts that wouldn’t in certain circumstances.  We need a course correction or everything becomes speech and there’s nothing special about a bookstore v. Target.  Result of very expansive theory: The things that people sell, like search results, are the result of expressive decisions; if those choices are expressive/organizing information, then that product is speech—which goes off the rails.
 
Q: Millian harm principle: w/o the 1A, the constraint is rational basis, which doesn’t require the harm principle. One way of understanding the 1A is as demanding something more than the harm principle (a certain kind of harm), or given the baseline of rational basis you could just demand the harm principle be satisfied.  Even if credit ratings are speech, you might have a harm-based justification to override it, in which case the Q about what justifies regulation of search engines would be based on protection, not just on coverage.  Compare the 4th Amendment context: analogies the SCt used fairly easily to justify things—video surveillance is like looking at people—are now at risk of abandonment (cellphone isn’t like other stuff in your pocket).  Disanalogies can also be recognized—leap in scope.
 
Whitney: The 4th Amendment comparison is a good one—resort to principles rather than analogies.
 
Simpson: true, might be covered but not protected.  For our purposes here, want to remain agnostic on the result of a protection decision within the scope of covered speech, because coverage decisions also have important implications for litigation.
 
Balkin: what search engines do more than anything else is serve democratic competence (Meiklejohn)—Meiklejohnian version of 1A would clearly lead to coverage for Google, leaving only protection remaining as the decision.  Many other algorithms would also pass Meikeljohn’s test for coverage.  Only an autonomy theory would say that only humans are bearers of speech, and distinguish between humans and their tools.
 
Simpson: we don’t think it’s as clear as that.  When I go to a search engine, what I think I’m seeing is a purely mechanically generated result.  In cases against search engines, the claim is a consciously gerrymandered result that is not that.  If Meikeljohn’s theory is about members of the demos having access to information they’d need in order to be participators, then the claim is that search engines, at least some of the time, distort exactly that information.
 
Balkin: Meiklejohn would never have said that b/c info you get is distorted, info is not protected. You need access so as to make your own judgments. Lots of information cooks the books. You’d have to argue that search engines have a different relationship to the public different to everyone else’s: information fiduciaries, with special duties to the public.  Grimmelmann, search engines as advisors.  Special duty by nature of service = ability to regulate in public interest; otherwise they’re in the same boat as any other info providers who cook the books (to mix a metaphor).  If a newspaper gives you a bunch of biased headlines, Meiklejohn has no problem with that. Only if an entity had a special duty to the public could it be regulated.
 
Corn-Revere: You’d have to reverse Tornillo for that.
 
Whitney: another possible move is new conceptions of autonomy/libertarian paternalism. Things that distort autonomy should not be unproblematically approved.
 
Balkin: they don’t try to apply nudges to First Amendment values.  Imagine a nudge to register all 18 year olds as Democrats.
 
Whitney: Democratic competence can have multiple meanings: people cannot always detect falsity/misleadingness. 
 
Balkin: but then you’re taking out a huge swatch of 1A doctrine.
 
Whitney: Accept that, though we are not arguing for that here.
 
Simpson: info fiduciary argument is worth pursuing: we’re trying to do more to theorize the special role that search engines have.
 
Sandy Baron: Q of responsibility for output in tort law.  Google doesn’t want to be responsible in that sense; can  you distinguish them in 1A protection/responsibility?
 
Whitney: It does seem there’s a tension for 1A protection for antitrust plus §230 protection because it’s not their speech. They’re neutral intermediary!  100% agree there’s an issue here. Identifying as a speaker is useful in some cases, harmful in others.
 
Q: Facebook isn’t the same as a search engine: trying to be objective/universal, presenting information as relevant. FB is more of a community; very different waters for the tech community.
 
Andrea Matwyshyn: Not everyone would agree w/that. 
 
Q: but FB will remove hate speech/terrorist content.

Wednesday, March 23, 2016

Inconclusive investigation isn't enough to give knowledge for contributory TM infringement purposes

Spy Phone Labs LLC v. Google Inc., No. 15-cv-03756-PSG, 2016 BL 86393 (N.D. Cal. Mar. 21, 2016)
 
SPL registered Spy Phone as a mark (for something, I presume), and submitted its Android app, SPY PHONE Phone Tracker, to Google.  It was downloaded over a million times, and Google took down a number of competing apps with similar names when SPL requested that it do so via Google’s online complaint form.  Whenever Google removed an app based on a trademark complaint, it sent the app developer a notification containing the complainant’s name and email address.
 
According to the complaint: The app stores information on phone locations and use on a secure server, and it periodically displays an icon to let users know the app is running; it’s free, but SPL generates revenue by running ads via Google on spyphone.com.  After one developer objected in May 2013, the relationship between Google and SPL eroded.  SPL submitted a trademark infringement complaint about another app, “Reptilicus.net Brutal Spy Phone,” and this time Google said it couldn’t determine the merits of the claim and refused to act.  Then in June, the Google Play team removed SPL’s own app, citing violations of Google's anti-spyware policy, allegedly despite the fact that SPL’s app complied with the policy.  Eventually, Google explained that although none of the functions of SPL’s app violated the anti-spyware policy, the name itself was unacceptable because it contained the word “spy.”  Though other apps used the same word, Google promised that it intended to prohibit all developers from doing so in the future.
 
Thus, SPL decided to drop its lawsuit and relaunch its app under the name “Phone Tracker.”  In October 2013, Google reinstated SPL’s developer account, but it deleted all the consumer reviews and records of downloads for SPL’s original app, and the new app managed only 260,000 downloads in ten months, leading to a steep reduction in SPL’s advertising revenue.29 Meanwhile, other apps continued to use “spy”  in their names with impunity.  So SPL resumed submitting complaints about other apps that used “spy” or “SPY PHONE” in their names, but instead of using the trademark infringement form—which would notify developers of SPL’s identity—SPL claimed violations of the same anti-spyware policy of which it had fallen afoul.  In July 2014, SPL made a complaint about another app from the Reptilicus.net developer, and Google again suspended SPL’s account without warning, this time for purportedly violating Google’s spam policy. SPL alleged that its app complied with that policy, but that the other developer had submitted a false complaint.  This other developer had multiple Play Store parental monitoring apps—which in itself violates Google's policies—and several of these apps contained the word “spy,” or even the phrase “Spy Phone,” in their names.  SPL allegedly received a letter from a “concerned” member of the Google Play team “confirming SPL’s suspicions.”  After this second takedown, Google searches for the phrase “spy phone” started to list competing apps before SPL’s website, and the top-listed result is now an app that allegedly infringes SPL’s trademark.
 
SPL sued the developer, who has not yet been served, and Google, alleging contributory trademark infringement against Google and tortious interference against all the defendants. The court first found that SPL didn’t allege facts sufficient to state a claim for contributory trademark infringement by Google. With an online marketplace, “a service provider must have more than a general knowledge or reason to know that its service is being used to sell counterfeit goods. Some contemporary knowledge of which particular listings are infringing or will infringe . . . is necessary.”  SPL didn’t allege that Google had notice for most of the apps; SPL intentionally made spyware complaints instead of trademark complaints in order to remain anonymous.  “But spyware complaints are not the same as trademark complaints, and Google could not be expected to respond to a complaint about one offense by investigating another.”
 
Google also didn’t ignore the initial Reptilicus.net app; it investigated and responded that it could not assess the merits of the claim.  The app might have used the words “Spy Phone” simply as a descriptor, as opposed to the distinctive prefix “Reptilicus.net Brutal,” and that fact would constitute a defense to infringement.  Uncertainty over the existence of infringement is relevant to an alleged contributory infringer’s knowledge.  That uncertainty, combined with Google’s investigation and response, made the allegations of knowledge implausible.
 
Nor did tortious interference work, because there was no allegation that Google committed any act wrongful apart from the interference itself.

Wednesday, November 04, 2015

Every Single Word, spoken no longer due to abusive copyright claim

Dylan Marron's "Every Single Word Spoken by a Person of Color in [X]" series, where X is a mainstream film, is a powerful indictment of popular culture.  Too powerful for Warner Brothers when it came to Gone With the Wind, apparently, as this notice indicates--and it's not a DMCA claim, either; Google went along with it.

Friday, October 16, 2015

Factor four: the most interesting part of the Google Books decision

Authors Guild v. Google, Inc., No. 13-4829-cv (2d Cir. Oct. 16, 2015)

Today's Google Books decision breaks no new ground in terms of transformativeness.  It is perhaps most interesting in its finessing of the question "what is a derivative work?" while suggesting that the answer may be a functional one based on similarity to the statutory list, as Pam Samuelson has argued in her own work. 

I was most struck by two related comments on factor four, which the court deemed the most important factor--perhaps a retreat from earlier post-Campbell cases--albeit one that had to be understood in light of factor one.  (The court didn't repeat the Second Circuit's previous statements that copyright owners can't create market harm by being willing to license transformative uses, but such repetition wasn't required.)  Though factor four is the most important, there are limits on its scope:
 
Lost sales must be substantial to make the fourth factor weigh against a transformative fair use:

But the possibility, or even the probability or certainty, of some loss of sales does not suffice to make the copy an effectively competing substitute that would tilt the weighty fourth factor in favor of the rights holder in the original. There must be a meaningful or significant effect “upon the potential market for or value of the copyrighted work.” 17 U.S.C. § 107(4).. . .
 
Also, lost sales must be based on protectable aspects of the work to count:

By entering “Roosevelt polio” in a Google Books search, the student would be taken to (among numerous sites) a snippet from page 31 of Richard Thayer Goldberg’s The Making of Franklin D. Roosevelt (1981), telling that the polio attack occurred in 1921. This would satisfy the searcher’s need for the book, eliminating any need to purchase it or acquire it from a library. But what the searcher derived from the snippet was a historical fact. Author Goldberg’s copyright does not extend to the facts communicated by his book. It protects only the author’s manner of expression.  Google would be entitled, without infringement of Goldberg’s copyright, to answer the student’s query about the year Roosevelt was afflicted, taking the information from Goldberg’s book. The fact that, in the case of the student’s snippet search, the information came embedded in three lines of Goldberg’s writing, which were superfluous to the searcher’s needs, would not change the taking of an unprotected fact into a copyright infringement.

Even if the snippet reveals some authorial expression, because of the brevity of a single snippet and the cumbersome, disjointed, and incomplete nature of the aggregation of snippets made available through snippet view, we think it would be a rare case in which the searcher’s interest in the protected aspect of the author’s work would be satisfied by what is available from snippet view, and rarer still—because of the cumbersome, disjointed, and incomplete nature of the aggregation of snippets made available through snippet view—that snippet view could provide a significant substitute for the purchase of the author’s book.

Tuesday, September 22, 2015

Google can't avoid class action just because damages are variable

Pulaski & Middleman, LLC v. Google, Inc., No. 12-16752 (9th Cir.  Sept. 21, 2015)
 
From 2004-08, many advertisers used Google’s AdWords to bid for Google to put their ads on websites. Pulaski sued under the California UCL and FAL, alleging that Google misled them as to the types of websites on which their advertisements could appear. Under Google’s auction system, the price of a click varied based on several factors: other bidders’ maximum bids, the ad’s quality score, and a “Smart Pricing” discount applied to the website where the ad was placed.  This discount comes from a ratio calculated by dividing the conversion rate for a lower-quality website by the conversion rate of google.com.  Although Google’s registration process only discussed Search Feed and Content Network placement—the first where the ads varied by the nature of the user’s search, and the second where the ads were determined by the content on the rest of the website, such as nytimes.com—AdWords also ran on parked domain pages and error pages.
 
After Google changed the AdWords program to let advertisers exclude parked domain and error pages, Pulaski sought only the equitable remedy of restitution. The district court held that common questions didn’t predominate because of damages issues, rejecting Yokoyama v. Midland National Life Insurance Co., 594 F.3d 1087 (9th Cir. 2010).  The court of appeals reversed.
 
Pulaski proposed three different methods for calculating restitution, all of which were based on a “but for” or “out-of-pocket loss” calculation: the difference between what advertisers actually paid and what they would have paid had Google informed them that their ads were being placed on parked domains and error pages.  One was based on the Smart Pricing discount ratio; a second method factored in the lower bidding that would have occurred had advertisers been allowed to bid separately on parked domains and error pages. A third method would provide full refunds for clicks on ads placed on parked domains and error pages. Each method would provide better results for different subgroups than the other methods.
 
The district court found that Rule 23(a) was satisfied in terms of numerosity, commonality, typicality, and adequate representation. However, for Rule 23(b)(3) predominance, the district court found that ascertaining who was even entitled to restitution would require individual inquiries, since “many advertisers … have no legal claim to restitution because they derived direct economic benefits from ads placed on parked domains and error pages.”  Individual questions would also arise in determining the amount of restitution owed.  Yokoyama held that damages calculations alone cannot defeat class certification, but that case involved a “workable method for calculating monetary recovery.” Here, the district court held, plaintiffs lacked such a workable method, given differing costs for each advertiser, each ad, and each click, overlaid with an auction process.
 
The court of appeals reversed. Entitlement to restitution was a common question.  The UCL and FAL are worded broadly; for false advertising claims, all a plaintiff must show is likely deception, not “individualized proof of deception, reliance and injury.” “[I]n effect, California has created what amounts to a conclusive presumption that when a defendant puts out tainted bait and a person sees it and bites, the defendant has caused an injury; restitution is the remedy.”
 
Google argued that class members were exposed to different information and thus couldn’t be treated together.  Pulaski responded that the conduct at issue was consistent: “all AdWords customers could select Search Feed pages, Content Network pages, or both; parked domain and error pages were never mentioned in AdWords’s sign-up materials; Google’s contracts with advertisers never disclosed that Google would place their ads on parked domains and error pages …; and Google’s materials answering frequently asked questions did not disclose ad placement on parked domain and error pages.”   Because Pulaski’s claims were based on the AdWords sign-up materials, “all of which were presented to putative class members through the same online portal,” Google’s argument failed.

The UCL and FAL provide for restitution as a remedy.  Courts need not make individual determinations about entitlement to restitution, which is available classwide once there’s a threshold showing of liability.
 
What about damage calculations?  Yokoyama held that “damage calculations alone cannot  defeat  certification.”  Google argued that Comcast Corp. v. Behrend, — U.S.—, 133 S. Ct. 1426 (2013), displayed Yokoyama, but that case involved a failure to provide a damages model that could link the plaintiff’s legal liability theory to only those damages attributable to that theory.  Comcast didn’t overrule Yokoyama as long as plaintiffs can show that their damages stemmed from the defendant’s actions that created the legal liability. Other circuits have agreed.
 
Google then argued that Pulaski’s proposed method for calculating restitution was “arbitrary,” but the court of appeals disagreed.  Restitution is “the return of the excess of what the plaintiff gave the defendant over the value of what the plaintiff received.” Restitution aims “to restore the defrauded party to the position he would have had absent the fraud,” and “to deny the fraudulent party any benefits, whether or not for[e]seeable, which derive from his wrongful act.”  In cases of deception, the consumer buys a product that he or she pays more for than he or she otherwise might have been willing to pay if the product had been labeled accurately.  This harm can occur even if the court views the products as functionally identical (as with two wines from different vintages), as long as the consumer wouldn’t.  Thus, UCL and FAL restitution is based on what a purchaser would have paid at the time of purchase had the purchaser received all the information.
 
California “requires only that some reasonable basis of computation of damages be used, and the damages may be computed even if the result reached is an approximation.” “[T]he fact that the amount of damage may not be susceptible of exact proof or may be uncertain, contingent or difficult of ascertainment does not bar recovery.”  Pulaski’s proposed method wasn’t arbitrary.  It didn’t need to account for post-purchase benefits, because the focus is the value of the service/what the consumer would have paid at the time of purchase. 
 
Pulaski’s main method of calculating restitution for Google’s placement of ads on lower-quality-than-expected web pages used Google’s Smart Pricing ratio, “which directly addresses Google’s alleged unfair practice by setting advertisers’ bids to the levels a rational advertiser would have bid if it had access to all of Google’s data about how ads perform on different websites.” Using a ratio that adjusts for web page quality “is both targeted to remedying the alleged harm and does not turn on individual circumstances.” The court therefore didn’t need to address the alternative methods.

Tuesday, May 19, 2015

Garcia v. Google reversed; many sigh in relief

Garcia v. Google, Inc., No. 12-57302, -- F.3d – (9th Cir. May 18, 2105) (en banc)
 
Judge McKeown wrote the majority opinion. Garcia was fooled into taking part in a movie that turned out to be the awful Innocence of Muslims, and received death threats because of it. She sought an injunction on copyright grounds, which was first denied, then granted, now lifted as the en banc court affirms the district court’s denial of an injunction.
 
The court began by describing Garcia’s burden as “doubly demanding: Because Garcia seeks a mandatory injunction, she must establish that the law and facts clearly favor her position, not simply that she is likely to succeed.” It’s a mandatory injunction because it requires Google to take affirmative action to remove, and keep removing, Innocence of Muslims from Google sites. Mandatory injunctions are disfavored except on strong showings of entitlement.
 
Starting with likely success on the merits: Garcia’s five-second “acting performance” didn’t give her a claim in a copyright. She might have publicity, defamation, or contract claims, but not copyright.
 
Copyright subsists “in original works of authorship fixed in any tangible medium of expression . . . [including] motion pictures.” Fixation must be done “by or under the authority of the author.” Here, Innocence of Muslims is a motion picture and a derivative work of the script. Garcia was the author of neither, but instead claimed separate protection for her five-second performance. The Copyright Office found that her performance was not a copyrightable work of its own, explaining that its “longstanding practices do not allow a copyright claim by an individual actor or actress in his or her performance contained within a motion picture.” For copyright purposes, “a motion picture is a single integrated work.” The court credited this expert opinion, which “reflects a ‘body of experience and informed judgment to which courts and litigants may properly resort for guidance.’”
 
By contrast, the dissent’s invocation of the Beijing Treaty on Audiovisual Performances was misplaced. First, the treaty is not in force; only six of the necessary thirty countries have ratified it. Second, though the US signed the treaty, the Senate has not ratified it. Though the PTO issued a fact sheet saying that performers’ performances were protected by copyright law, the PTO lacks legal authority to interpret the Copyright Act. (RT: Also, it may still be the case that primary performances make performers co-authors or even sole authors where they control the recording—Garcia didn’t rely on joint authorship or sole authorship claims as to the film as a whole.)
 
Aalmuhammed v. Lee, 202 F.3d 1227 (9th Cir. 2000), was a useful starting point; the meaning of “work” is the first step in analyzing joint authorship.  When a work is “prepared by two or more authors with the intention that their contributions be merged into inseparable or interdependent parts of a unitary whole,” the work becomes a “joint work.” But Garcia disclaimed joint authorship of the film.  Defining a “work” based upon “some minimal level of creativity or originality . . . would be too broad and indeterminate to be useful.”  Aalmuhammed’s animating concern was that a broad definition of “work” “would fragment copyright protection for the unitary film Malcolm X into many little pieces.” We shouldn’t splinter a film into many different “works” in the absence of independent fixation. “Treating every acting performance as an independent work would not only be a logistical and financial nightmare, it would turn cast of thousands into a new mantra: copyright of thousands.”
 
The dissent’s hypotheticals could be addressed with careful application of statutory definitions.  (The IP profs are still going at this, but I’m with the majority here.)  The dissent’s reliance on Effects Associates, Inc. v. Cohen, 908 F.2d 555 (9th Cir. 1990) ([pointedly] Kozinski, J.), was inapposite, because no one in that case challenged the existence of an independent copyright on the special-effects footage that was independently fixed. 

Contracts and WFH govern “much of the big-budget Hollywood performance and production world,” and implied licenses often take care of the rest (the district court found one here). “But these legal niceties do not necessarily dictate whether something is protected by copyright, and licensing has its limitations. As filmmakers warn, low-budget films rarely use licenses.” Plus, contracts aren’t perfect, especially for third-party distributors like YouTube and Netflix, who won’t have easy access to any licenses; “litigants may dispute their terms and scope; and actors and other content contributors can terminate licenses after thirty five years. Untangling the complex, difficult-to-access, and often phantom chain of title to tens, hundreds, or even thousands of standalone copyrights is a task that could tie the distribution chain in knots.”  And filming group scenes “like a public parade, or the 1963 March on Washington, would pose a huge burden if each of the thousands of marchers could claim an independent copyright.”
 
In a footnote, the court of appeals found that the district court’s implied license finding wasn’t clearly erroneous.  “Although Garcia asked Youssef about Desert Warrior’s content, she in no way conditioned the use of her performance on Youssef’s representations.”
 
In addition, the court held, Garcia didn’t fix her acting performance in a tangible medium.  She wasn’t the one who did the fixation.  (The court noted the Copyright Office’s distinction between acting performances “intended to be an inseparable part of an integrated film” and “standalone works that are separately fixed and incorporated into a film,” the latter of which could be protected by copyright.)  Youssef and his crew did the fixation.  And Garcia claimed that she never agreed to the film’s ultimate rendition “so she can hardly argue that the film or her cameo in it was fixed ‘by or under [her] authority.’”
 
The district court didn’t err, and the law didn’t clearly favor Garcia.  But just in case, the court went on to address irreparable harm, because of the alleged threat to Garcia’s life. “Garcia understandably takes seriously the fatwa and threats against her and her family, and so do we.”  The problem was that there was no match between these risks and her substantive copyright claim. 
 
For irreparable harm in a copyright claim, as opposed to a fraud or false light claim, her harm needed to be “harm to her legal interests as an author.”  Copyright is supposed to promote progress and create an incentive to disseminate works; its justification is the protection of an author’s commercial interest, not the protection of secrecy.  Garcia’s “severe emotional distress, the destruction of her career and reputation,” and receipt of credible death threats, were “untethered from—and incompatible with—copyright and copyright’s function as the engine of expression.”  Copyright’s function is not to protect privacy or to guard against emotional distress: “such damages are unrelated to the value and marketability of [authors’] works.” US law doesn’t provide a right to be forgotten or moral rights. “We do not foreclose that in a different circumstance with a strong copyright claim, a court could consider collateral consequences as part of its irreparable harm analysis and remedy. But such a case is not before us.”
 
In addition, Garcia waited months to seek an injunction, and it wasn’t an abuse of discretion to weigh this delay undercut her irreparable harm claim.  Garcia acted once the film was translated into Arabic and sparked death threats against her. “But that proves the point: the gravamen of Garcia’s harm is untethered from her commercial interests as a performer, and instead focuses on the personal pain caused by her association with the film.”
 
The panel’s injunction, which ordered Google to take down all copies of the film from all platforms under its control, and take “all reasonable steps” to prevent further uploads, was wrongly granted.  Even in its amended form, which allowed the posting of any version of the film without Garcia’s performance, was wrong, putting the court “in the uneasy role of film editor.”  It would only matter if Google (or someone else) decided to edit another’s copyrighted film.  “To no one’s surprise, the end result was the same: the entire film remained removed from YouTube.”  Not only was this order incorrect as a matter of law, it disserved the First Amendment by suppressing a politically significant film.  Copyright isn’t categorically immune from First Amendment challenge, and the takedown order directed at “a film of substantial interest to the public” was a classic prior restraint.  A thin copyright claim was no warrant for such action.
 
Judge Watford concurred, and would have ruled solely on irreparable harm, reserving touchy copyright issues for later.  He would have accepted that the risk of death qualified as irreparable injury, but concluded that Garcia couldn’t prove a causal connection between the irreparable injury and the conduct she sought to enjoin: she couldn’t show that removing the film from YouTube would materially reduce the risk of death she faced. “Garcia is subject to the fatwa because of her role in making the film, not because the film is available on YouTube.”  Correcting misperceptions about her role in the film—she was duped into participating, she never said the offensive words her character speaks in the film, and she strongly opposed the film’s message—could help, but she’s already done everything within her power to dissociate herself from the film.  Garcia’s expert on Islamic and Middle Eastern law didn’t claim that removing the film from YouTube would likely cause the fatwa against her to be lifted, but merely opined: “If she is successful in pulling the content down from the internet, it will likely help her in terms of believability of her message condemning the film and its message.” That was too little in the way of evidence of likely impact on the key audience, especially given that demanding the takedown alone spoke of her sincerity.
 
Judge Kozinski dissented. Garcia’s dramatic performance had the minimal creativity to be copyrightable subject matter, it was original, and it was fixed at the moment it was recorded.  That’s enough.  [That’s enough to make it at least part of a work … but who’s the author of that work?]  The majority says that Garcia’s performance wasn’t a work, “apparently because it was created during the production of a later-assembled film. But if you say something is not a work, it means that it isn’t copyrightable by anyone.” So the majority’s definition of work means that no one has a copyright in any part of Garcia’s performance, “even though it was recorded several months before Innocence of Muslims was assembled.” If only the film is a work, then the copyrightability of “vast swaths of material created during production of a film or other composite work” is in doubt, such as every take that doesn’t become part of a final movie. “If some dastardly crew member were to run off with a copy of the Battle of Morannon [from Lord of the Rings], the dastard would be free to display it for profit until it was made part of the final movie. And, of course, the take-outs, the alternative scenes, the special effects never used, all of those things would be fair game because none of these things would be ‘works’ under the majority’s definition.”  The same for draft chapters of books, or parts thereof.
 
RT: There may be unusual cases, but most takes, for example, should be at least derivative works or reproductions of the script, which is copyrightable, so I don’t see a practical problem developing. The question is which hypotheticals you find most troubling/plausible, and I think the risk of making every five seconds into its own work—with corresponding effects on substantial similarity and fair use analysis—justifies taking the risks Kozinski identifies instead.
 
Effects Associates dictates a contrary result, because it held that a special effects company retained a copyright interest in its footage even though it became part of the film.  Like the special effects shots, Garcia’s performance was also “separately fixed and incorporated into” Innocence of Muslims.  Youssef could’ve sold the clip featuring Garcia to someone else; it might not have had much value, but then neither did the special effects from The Stuff. 
 
Here I think Kozinski cleverly equivocates between “Garcia’s performance” and “the footage.”  Garcia’s performance was not separately fixed from the other elements that made up the movie that were in the shots with her; it was fixed along with them.  That wasn’t true of the distinct shots created in a different place and at a different time in Effects Associates.  If (as Kozinski ten suggests) the effects had been later added into a greenscreened performance where the actors were staring at tennis balls, I think you’d get the Garcia result and not the Effects Associates result.  I don’t think we’ll ever get a perfectly pure rule here, but that’s copyright for you.
 
This isn’t an unfamiliar question. Statutory damages cases ask us to identify how many “works” are at issue, and while the statute provides an answer for compilations that are infringed, it doesn’t provide all the answers.  For example: suppose someone makes a for-profit, nontransformative, non-fair use Batman movie—alley, gunshot, pearls falling into a puddle, all that jazz.  I take it we wouldn’t want to say that the infringer had infringed each of the gazillion registered works that make up the Batman corpus and recount their own versions of this story.  What the infringer has done is infringe the Batman story (we can also count characters, but then we have to ask whether the characters are registered; anyway, counting by number of characters is a very different thing than counting by number of registered works).  By contrast, someone who reproduces the Dark Knight movies has copied specific works and the statutory damages should be calculated that way (and not as an infringement of all the other registered works that also tell the Batman story).  Of course we get slightly different definitions of “works” from the different situations, because we have different concerns about the interests invaded. But I don’t think that makes the whole edifice illegitimate.
 
Anyhow, back to the dissent: the majority erred in saying that Garcia couldn’t be an author because she didn’t play a role in the fixation.  A work is fixed when the fixation is by or under the authority of the author.  She’s an author because without her, there was only a script.  Her performance had at least “some minimal degree of creativity.” “To dispute this is to claim that Gone With the Wind would be the same movie if Rhett Butler were played by Peter Lorre.” (RT: Okay, that’s overblown. Innocence of Muslims is surely no GWTW, and Garcia’s 5-second character no Rhett Butler.  Actually, now that I’m thinking about this, isn’t Kozinski’s argument here a workaround for the rule that a character has to be sufficiently defined to be protected by copyright?  After all, Garcia’s physical characteristics aren’t “creative” or “original” in the copyright sense, so it’s just what she does in those five seconds that he argues makes her an author.)

Because Garcia didn’t sign away her rights, she acquired a copyright in her performance the moment it was fixed. “It’s not our job to take away from performers rights Congress gave them.” The Register of Copyrights was a member of the U.S. delegation that signed the Beijing Treaty on Audiovisual Performances, which would recognize Garcia’s rights in her performance because it provides that “performers” have the “exclusive right of authorizing . . . the fixation of their unfixed performances,” and “reproduction of their performances fixed in audiovisual fixations, in any manner or form.”  The PTO said that U.S. law was “generally compatible” with the Treaty, as “actors and musicians are considered to be ‘authors’ of their performances providing them with copyright rights.” The Copyright Office surely wouldn’t have signed on if it believed that the Treaty’s key provisions were inconsistent with U.S. copyright law. (Surely! After all, US negotiators are never captured and never engage in policy laundering of that type.)
 
The Copyright Office said “an actor or an actress in a motion picture is either a joint author in the entire work or, as most often is the case, is not an author at all by virtue of a work made for hire agreement.” But Garcia isn’t in either category, so she must be an author.  (It’s the “so” where we part company.  This really should be fixed by a better understanding of joint authorship.)
 
There’s not much risk of harm from Kozinski’s position, since “[t]he vast majority of copyright claims by performers in their contributions are defeated by a contract and the work for hire doctrine. And most of the performers that fall through the cracks would be found to have given an implied license to the film’s producers to use the contribution in the ultimate film.”  The remaining few wouldn’t be able to claim the “valuable” rights of joint authorship. (RT: Yeah, only total ownership of a chunk!  That doesn't pose any problems at all!) That’s why Effects Associates didn’t cause filmmaking to grind to a halt. Anyway, the Supreme Court said in Tasini that “speculation about future harms is no basis for [courts] to shrink authorial rights,” and neither the predicted database disasters or the disasters predicted by amici in CCNV took place.  (Actually, the Tasini situation is rather more complicated than that, and the case just ended after two decades—but at least that was susceptible to class treatment!)
 
The risks that this rule imposes on producers and ISPs is for them to fix by contract.  “Google makes oodles of dollars by enabling its users to upload almost any video without pre-screening for potential copyright infringement. Google’s business model, like that of the database owners in Tasini, assumes the risk that a user’s upload infringes someone else’s copyright, and that it may have to take corrective action if a copyright holder comes forward.”  (Takeaway: Kozinski really, really doesn’t like ISPs.)
 
Doomsday claims are overstated: “our injunction has been in place for over a year; reports of the internet’s demise have been greatly exaggerated.” Garcia’s copyright claim was likely to succeed and she made an ample showing of irreparable harm. “It’s her life that’s at stake.”

Wednesday, April 01, 2015

ANA conference: Miss. AG Jim Hood

Keynote Address, Jim Hood, Attorney General, State of Mississippi, President, National Association of Attorneys General (NAAG)
 
Law enforcement has to be moving to the internet, where crime is going. Worked with ISPs on child porn/hash screening. We do hacking, extortion, data breach, IP theft.  He fears in our lifetime a hacker will shut down our electric grids. Survived Katrina; carried a gun the first few weeks thereafter.  Law-abiding people will steal water, gasoline, whatever they need to protect their families if the electrical grid is shut down.
 
AGs have multistate working groups. If you have a data breach, and call my office as GC, I have a direct link to that company.  If you hire a large firm with a data breach practice, most AGs will think “this is the defense firm we have to deal with” and their eyes glaze over. Work with the AG instead.  If you hire a defense firm, we assume you’re guilty, that’s our law enforcement action.  BP oil spill: initially well handled, GC reached out; then they figured out they were really in trouble.  If you have a small data breach, work with AG’s office. Hard to point a finger when we may not have protected our own gov’t systems as we should.  Compare to your competitors—big box store may need different procedures than a bank. AGs are putting together a manual with suggestions; not necessarily best practices.  Primarily for small businesses suffering a data breach.
 
Most AGs want federal legislation to be as stringent as the most stringent state legislation. Probably will see some movement in the next year or two.  We fight federal preemption every day, Democrats and Republicans.  There is a lot of overreach—federal district judges in particular haven’t respected state sovereignty as they should.  (Heh.)  CAFA: 49 AGs said “exempt AG actions from this act,” and Senators said they wouldn’t be covered, but the 5th Circuit said that AGs were covered anyway; SCt reversed 9-0. There’s a constant battle.  But we’ll probably reach agreement on data breach.
 
We just appealed the Google decision.  The Sony hack: I didn’t have a clue they knew Mississippi had an AG; I didn’t know anything about operation Goliath or any of this stuff.  They acted like the movie theaters got an AG to go after them, but I’ve been doing it for 6-8 years. Motion picture industry has been involved since the 1920s when people stole film off trucks. Counterfeit items hurt our consumers.  Drugs, etc. 
 
After Google entered into a plea w/fed. gov’t, paid $500 million to avoid conviction, b/c DOJ found that marketing division was getting around their system to allow Canadian pharmacies to advertise—really dope dealers in South America and Russia. Google folded b/c DOJ got their emails.  Complaints from parents: investigators made buys online, like oxycodone. Autocomplete used to complete “prescription” with “buy prescription drugs without prescription.”  We bought from their ads.  Monetizing the sale of illegal drugs is a problem.  YouTube videos would say “here’s how to buy drugs w/o a prescription.”  (§230?) 
 
IP theft is important—Mississippi has an inordinate number of artists and writers—but the AGs weren’t badgering Google out of nowhere.  We bought pain and birth control pills; often we got ripped off, or got counterfeits—during Google’s 2 year probationary period and we provided that information but the fed gov’t hasn’t done anything so far. Engaging on these tech issues is not new.  3-D printers are an emerging issue.  Will change business more than the internet. Will be able to print a cellphone. Won’t need Chinese labor. Also IP theft.  Guns, drugs.
 
Encryption: will you get sued if you don’t encrypt? You can’t encrypt w/in business, but if you send it to someone else, they need to decrypt.  Apple iOS: Now we have digital info in almost every criminal case—texting, photos, emails. We rescue kids by working with Facebook: we find them if they’ve run away. FB works with us.  Law enforcement access saves lives.  Crooks want privacy protection—helps sell phones.  (So much for data breach concerns.)   If we can’t get cellphone companies to keep some code for us, we need a law to give us the best evidence.  What if a guy gets shot and the video on the phone is the only evidence?
 
Q: in the past, NAG has cooperated w/FTC. What’s it like now, and how do you see it going forward?  What are the priorities in consumer protection, and do they match those of the FTC?
 
A: have a good relationship w/FCC and FTC.  Antitrust can be slow, but generally they deal with a bigger bureaucracy.  Ads to kids: higher standards. Particularly online, all the apps targeted at kids.
 
Q: why aren’t you boycotting Indiana? Aren’t civil rights a critical issue as well?
 
A: Governor of Conn. banned state funds for travel there.  They’re going to have to deal with their own problem.
 
Q: distinguishing between ads and content—are the AGs involved?
 
A: not as much, but in the area of Google ads/AdWords, when we see “how to murder your wife” for a YouTube video, we notice that.  Have to disclose where you place ad.  We will look at putting legit ads beside illegit ads. 
 
Q: self-regulation, including on piracy: ecosystem challenge is where the liability should be put.  There’s a challenge in putting liability on intermediaries.  We’ve come out on having advertisers do contracts and push down limitations on ads delivered on pirate sites.
 
A: companies have a role here. Talk to the AGs but AGs can’t carry all your water.  A lot of times competitors come to us. We try to stay out of individual class actions.