Technology & Marketing Law Blog

Court Denies Restraining Order Against Ex-Boyfriend Who Threatened to Post Revenge Porn -- EC v. CBT

E.C. v. C.B.T., SR., A-1185-12T2 (N.J. Ct. App. May 6, 2013)

Plaintiff and defendant lived together between May and August 2011, and after that had a relationship that plaintiff characterized as dysfunctional. Plaintiff sought and obtained a restraining order based on defendant’s stalking. According to her, defendant:

- caused a scene at her workplace;

- threatened to tell her parents that the father had cheated on her stepmother;

- told her he planned to post compromising photos of her on craigslist;

- confronted her at a bar after her high school reunion and followed her and a male friend of hers to the friend’s house where he argued with the friend;

- in the days after the in-person confrontation, plaintiff and defendant exchanged text messages and other communications.

After these incidents, plaintiff continued to communicate with defendant, but ultimately broke it off. Prompted by a message defendant posted to her friend’s Facebook page (the content of which was not discussed by the court), plaintiff sought and obtained a restraining order.

Defendant had a different view of the events. He said that they had an “on-again/off-again” relationship. Regarding the reunion, he said the plan was for her to attend the reunion and for them to hook up after. He apparently went to the bar which was close to his house, saw plaintiff, and was concerned about whether plaintiff would be able to drive home. He said he followed her and the friend to the friend’s house where the friend started a scene. He said he couldn’t recall ever insulting or threatening plaintiff via email or text message, but he testified that he regularly exchanged messages with plaintiff.

The trial judge didn’t give much credence to the online activity at all:

I don't give a lot of credence to e-mails and Facebook and all that nonsense because that's not a face to face exchange. Nobody is in danger. Nobody suffers from that.

The judge was also confounded about the risqué pictures and why plaintiff sent them to defendant in the first place:

She tells him to stop and then submits naked pictures to him. That's the dilemma I have. What is she--what is she trying to express by doing that? Is that leave me alone? Is that I--the relationship will continue? I don't know. I can't understand that motive--motivation.

Nevertheless, the court granted the restraining order on the basis that defendant caused plaintiff fear by appearing at the bar and following her.

The appeals court reverses, finding that issuance of a stalking-based restraining order requires two elements--plaintiff must prove (1) a predicate act (in this case, stalking); and (2) that an order is necessary to protect the plaintiff from immediate danger or other acts of domestic violence. The court notes that the trial judge was equivocal in her own findings of fact, and failed to explain the many inconsistencies in plaintiffs’ testimony (and ignored defendant’s allegations). The trial judge also failed to find a “course of conduct,” which is required for stalking. The court does say that defendant’s conduct was “immature and jealous,” but there was insufficient evidence to find that he engaged in stalking.

The court also says that the judge failed to make a finding that the order was necessary to protect the plaintiff’s safety or person, pointing out that the trial judge also did not explain the equivocal nature of plaintiff’s conduct (i.e., she appeared to interact with him, despite her testimony that she was fearful of contact with him).
__

New Jersey's anti-harassment statute has a broad reach, which at least in one instance, has been curtailed by an appeals court. ("New Jersey Appeals Court Reverses Anti-Harassment Order Based on Emails – E.L. v. R.L.M.") This case is in a similar vein, and it illustrates how confusing the statute is in application.

The elephant in the room of course is defendant's threat to make public the risqué pictures that plaintiff sent to defendant. (The trial judge strangely brushed aside the online activities altogether.) Although the opinion is unclear on this issue, it's possible that this is what prompted plaintiff to seek the order in the first place. In any event, New Jersey's anti-stalking statute could conceivably reach this activity, except that the appeals court found no course of conduct to begin with.

The trial judge's comments question why plaintiff sent the photos after things appeared to be less-than-stable in the relationship. This is starting to go down the path of blaming the victim. Irrespective of when she sent the photos to the defendant, what's the plaintiff supposed to do now? Waiting until the photos are published is untenable, as the horse would be out of the barn, and no one has a practical solution to cleaning up images from the internet. On the other hand, there are clearly speech concerns with applying an overly broad law such as NJ's stalking law to solve this problem.

[photo credit: Shutterstock/Memo Angeles - "Internet Troll Using a Computer"]

Talk on Designing Optimal Safe Harbors and Immunities

By Eric Goldman

Recently, I posted a Forbes article entitled "Designing Optimal Safe Harbors and immunities." Based on the feedback I got to the article, I plan to convert it into a full-blown law review article eventually (i.e., it could take me years). Because I plan to invest more into this project, I workshopped the paper at the University of Washington (thanks Zahr for organizing this!). From the talk: my presentation slides, audio recording (item #36) and video recording (item #37).

Google Crushes Home Decor Center’s Trademark Challenge to AdWords

By Eric Goldman

Home Decor Center, Inc. v. Google, Inc., 2:12-cv-05706-GW-SH (C.D. Cal. May 9, 2013)

Home Decor Center sued Google in the wake of the Fourth Circuit's Rosetta Stone ruling. The lawsuit did not go well for it. Recently, the court granted Google’s summary judgment motion on two grounds. The trademark challenges were dismissed because the court held that “Home Decor Center” was generic and therefore ineligible for trademark protection. All of the non-trademark claims were dismissed per 47 USC 230. Thus, not only has Home Decor Center’s lawsuit died, so has its trademark. The only thing that could be worse is if the court orders Home Decor Center to pay Google’s legal fees (I imagine Google will ask). Perhaps Home Decor Center will appeal the ruling—at this point, it has nothing to lose in the case other than more legal fees.

Purists will be disappointed that yet another court sidestepped the conceptually interesting question of when Google could be liable for selling trademarks as keyword ad triggers—a question that still has not been definitively resolved in court after more than a decade of legal battles, and now appears that it may never be answered. Google recently settled the CYBERsitter lawsuit, so Google has whittled its AdWords trademark litigation docket down to just two pending lawsuits: Parts.com and Ison. Parts.com will almost certainly end like this one did—a complete plaintiff loss on the merits plus a finding that the putative trademark Parts.com is generic—and Ison has already lost at the lower court and has zero chance on appeal. Thus, with just a little mop-up work remaining, Google has effectively cleared its AdWords trademark docket, reinforcing that Google has won the trademark battles over AdWords. It will be interesting to see if any future trademark owners make the sucker’s bet of challenging the Mighty Google.

Home Decor Center is Generic

The Home Decor Center lawsuit baffled me generally, but perhaps the most baffling fact is that it only had a trademark registration on the Supplemental Register, not the Principal Register. The Supplemental Register is only for descriptive trademarks that haven’t achieved secondary meaning, so the registration is usually worthless in court. Indeed, though the PTO makes a finding that the term is a descriptive trademark and not generic as part of granting a supplemental registration, courts don’t need to defer to the PTO’s determination, and this court didn't.

To help crush the trademark, Google conducted at least two consumer surveys, while Home Decor Center apparently didn’t conduct any. Trademark defendants don’t typically do their own consumer surveys about the plaintiff's trademark, but Google's surveys proved devastating to Home Decor Center’s case. I don’t know how much money Google spent on its consumer surveys, but I doubt the price tag was less than a quarter-million dollars. Thus, Google was willing to invest in a nuclear flyswatter to destroy a lawsuit it was likely to win even without the surveys. In a similar vein, Google did a consumer survey in the Ison case, which was an even weaker lawsuit. Google's willingness to vastly outspend its opponents provides a good cautionary tale for the next plaintiff who thinks it can challenge Adwords “on the cheap.”

The court summarized its discussion on genericness:

Combining the words "home decor" with "center" does not primarily denote a specific origin or source. Rather, it merely identifies the "type of product" sold by Plaintiff, answering the consumer's "what are you" question with the response, a "center" that provides "home decor."

Overall, I’ve previously argued that the term “[noun] Store” is categorically generic for a retailer selling [noun]. This ruling is similar; using “center” rather than “store” doesn't really change the analysis. As Parts.com will soon find out, “[noun].com” is also generic for an online retailer of [noun].

The Section 230 Analysis

The court says that Home Decor Center's state law claims are all derivative of its trademark claims. Thus, when the trademark died, so did the derivative claims. But even if that wasn't true, the court says the derivative claims are preempted by 47 USC 230. This is because Home Depot, the AdWords advertiser who purchased the allegedly problematic ads, used Google's tool to dynamically insert the search query (in other words, the purported trademark) into the ad copy. When the advertiser uses Google's dynamic keyword insertion feature, Google doesn't "create or develop" the ads in question (with cites to Jurin and Goddard). Home Decor Center presented some evidence that Google's sales rep discussed the ad copy with Home Depot, but ultimately it appears Home Depot wrote all of the ad copy itself. Compare the murky Section 230 discussion in the CYBERsitter case, now effectively orphaned by settlement of that case.
___

For more on this topic, see my recent blog posts on trademark infringement and keyword advertising:

* Suing Over Keyword Advertising Is A Bad Business Decision For Trademark Owners
* Florida Proposes to Ban Competitive Keyword Advertising by Lawyers
* More Confirmation That Google Has Won the AdWords Trademark Battles Worldwide
* Google's Search Suggestions Don't Violate Wisconsin Publicity Rights Law
* Amazon's Merchandising of Its Search Results Doesn't Violate Trademark Law
* Buying Keyword Ads on People's Names Doesn't Violate Their Publicity Rights
* With Its Australian Court Victory, Google Moves Closer to Legitimizing Keyword Advertising Globally
* Yet Another Ruling That Competitive Keyword Ad Lawsuits Are Stupid--Louisiana Pacific v. James Hardie
* Another Google AdWords Advertiser Defeats Trademark Infringement Lawsuit
* With Rosetta Stone Settlement, Google Gets Closer to Legitimizing Billions of AdWords Revenue
* Google Defeats Trademark Challenge to Its AdWords Service
* Newly Released Consumer Survey Indicates that Legal Concerns About Competitive Keyword Advertising Are Overblown

[Photo Credit: Give Pests The Flick With An Isolated Hand Flicking A Huge Dead Fly // ShutterStock]

Suing Over Keyword Advertising Is A Bad Business Decision For Trademark Owners--General Steel v. Chumley (Forbes Cross-Post)

By Eric Goldman

General Steel Domestic Sales, LLC v. Chumley, 2013 WL 1900562 (D. Colo. May 7, 2013)

Trademark owners rarely win keyword advertising lawsuits in court. Reinforcing this conclusion, another trademark owner lost a trial over competitive keyword advertising despite a number of key facts in its favor. Given how often trademark owners lose keyword advertising lawsuits, why do they keep wasting their time and money?

The Lawsuit

General Steel Corporation and Armstrong Steel Corporation compete in the "prefabricated steel building business." Chumley, a former employee of General Steel, launched Armstrong Steel and ran Google ($GOOG) AdWords ads on the keyword "General Steel" with ad copy such as:

General Steel Buildings www.ArmstrongSteelBuildings.com Price Your Building Online Or Let Us Do It. Guaranteed Lowest Prices!

and

General Steel Buildings Price an Armstrong Steel Building Online in Minutes Or Let Us Do It. www.ArmstrongSteelBuildings.com.

(As usual, the court doesn't discuss whether Armstrong Steel broad-matched the word "steel").

Eventually, Armstrong Steel made its ad campaign more clearly comparative, buying the keyword “General steel buildings” to display ad copy such as:

Don’t Buy General Steel Without Pricing Armstrong First. Price a Steel Building in Minutes! www.ArmstrongSteelBuildings.com

In addition to the keyword advertising campaign, Chumley (or an employee of his) issued press releases that contained false claims and were attributed to a fictitious employee, and the defendant's website made additional false claims. Chumley further falsely published Internet postings in the name of General Steel's CEO. Rebecca Tushnet discusses the false advertising angles of the case.

General Steel sued Armstrong Steel and Chumley for false advertising and trademark infringement. After a bench trial, the judge ruled for General Steel on the false advertising claims and for the defendants on the trademark claims.

Regarding keyword advertising, General Steel's trademark claims failed because it couldn't show actionable consumer confusion. The court says it "suspects" that some searchers clicked on Armstrong Steel's ads believing it would take them to General Steel's website. However, that wasn't enough to win the case (especially after a key Ninth Circuit 2011 ruling on keyword advertising) because "[o]ther than its suspicions, plaintiff offered no evidence at trial of actual confusion," and because steel buildings are complex purchases that buyers will research carefully. The court concludes:

While Armstrong was using the term to refer to plaintiff’s company, the Court does not find the record supports the finding that such use was likely to cause confusion among consumers in light of all of the surrounding information that identified Armstrong Steel as the source of the website and distinguished Armstrong Steel from General Steel.

The court also rejects a common argument by trademark owners that "potential customers entering the term 'general steel' into a search engine are searching exclusively for that company, as opposed to executing a broader search for all companies selling similar products." (See the uncited study by Franklyn and Hyman supporting this analysis). The court instead says "the connection between the search term entered and the appearance of an advertisement is too attenuated to suggest an actual affiliation between the two."

Implications

This is really a devastating loss for the plaintiff--and for trademark owners generally. On its face, this case looked like a sure win for the trademark owner. After all, the ad copy referenced the trademark in a non-comparative way--usually thought to be a "no-no"--plus the advertiser engaged in false advertising elsewhere. If a trademark owner can't win a case with those attributes, when can it win a competitive keyword advertising case?

As it turns out, the answer is "rarely." Though not a lot of cases have reached the final merits of trademark claims over keyword advertising, the rulings (summarized below) demonstrate why trademark owners should think carefully before suing over competitive keyword advertising:

* plaintiff got injunction: CJ Products v. Snuggly Plushez (2011); InternetShops v. Six C (2011) (note: defendant admitted trademark liability, so the opinion only deals with remedies).

* plaintiff won summary judgment: Storus v. Aroa (2008).

* plaintiff won at trial: Binder v. Disability Group (2011). This case was decided before the Ninth Circuit ruling in Network Automation, and I believe it’s no longer good law.

* defendant won summary judgment: J.G. Wentworth v. Settlement Funding (2007); Designer Skin v. S&L Vitamins (2008); 1-800 Contacts v. Lens.com (2010); Montana Camo v. Cabela’s (2011); Jurin v. Google (2012) (note: unlike the other cases, in Jurin the defendant was Google, not the advertiser); CollegeSource v. AcademyOne (2012).

* defendant won at trial: Fair Isaac v. Experian (2009) (technically, the final win came in a post-trial ruling after a jury trial); College Network v. Moore (jury ruling in 2009; affirmed on appeal in 2010); Consumerinfo v. One Techs. (2011) (jury trial); General Steel v. Chumley (bench trial).

For more on this topic, see my recent blog posts on trademark infringement and keyword advertising:

[Photo Credit: Closed Road // ShutterStock]

Judge’s Facebook Friendship With Victim’s Parent Does not Taint Proceeding -- Youkers v. Texas

[Post by Venkat Balasubramani]

Youkers v . Texas, No. 05-11-01407-CR (Tx. Ct. App. May 15, 2013) [pdf]

Youkers was convicted for tampering with evidence after he was indicted for assaulting his girlfriend who was pregnant with his child. He entered into a plea deal under which his prison sentence would be suspended and he would have to pay a fine. Three months after the deal, the State filed a motion to revoke the suspended sentence (and send Youkers to prison), contending that he violated the terms of his supervision.

Younkers entered an “open plea of true” and sought leniency on the basis that while he did not previously have a stable place to live, he did now. The trial judge rejected his contentions and sentenced him to 8 years in prison. The judge also rejected his request for a new trial.

Youkers appealed, and raised (among other issues) the fact that (1) the trial judge was Facebook friends with the victim’s father and (2) in the context of the initial proceeding, the victim’s father sent the trial judge an ex parte communication in the form of a Facebook message. The appeals court says none of this rises to the level of improper bias.

The Communications Do not Show Bias: Youkers' appeal was focused on the Facebook friendship, but the court points out that the communications between the victim’s father and the judge in the initial proceedings do not show bias. The communication took place in the initial proceeding, was actually favorable to Youkers (it sought leniency), and the trial judge treated it as an ex parte communication. (The judge placed the communication in the record, and warned the father that such communications were not allowed.)

Mere Fact of Facebook Friendship Also Does not Show Bias: Turning to the key question of whether merely being Facebook friends is sufficient to show bias, the court says no. Citing to a recent ABA Standing Committee report, the court says that judges are not prohibited from using social media:

Allowing judges to use Facebook and other social media is also consistent with the premise that judges do not forfeit [their] right to associate with [their] friends and acquaintances nor [are they] condemned to live the life of a hermit.

In fact, such a regime "would . . . lessen the effectiveness of the judicial officer." (Citing Comm. on Jud. Ethics, State Bar of Tex., Op. 39 (1978).) Nevertheless, while social media may be useful professionally (judges are elected, and the court recognizes that social media is part of any modern day judicial campaign) and personally, judges must still abide by the applicable ethical rules.

The court looks to the Texas Code of Judicial Conduct and several Canons, all of which prohibit any actions that undermine public confidence in the judiciary or that would convey anything other than impartiality. On the core question of where mere Facebook friendship warrants recusal, the court says no:

Merely designating someone as a "friend" on Facebook "does not show the degree or intensity of a judge's relationship with a person." ABA Op. 462. One cannot say, based on this designation alone, whether the judge and the "friend" have met; are acquaintances that have met only once; are former business acquaintances; or have some deeper, more meaningful relationship. Thus, the designation, standing alone, provides no insight into the nature of the relationship.

Youkers was required to produce additional evidence that there was an improper relationship and he failed to do this. The court does say that judges should be careful to not let third parties (i.e., the victim’s father in this case) convey the impression that they have special sway over the judge, but the judge did the right thing here. As soon as the father made a statement, the judge flagged it as an improper ex parte communication and instructed the father to not send any more messages along these lines.

Shocker of shockers. For the most part, courts recognize that judges are also humans, and things like social networks should not be off-limits to them merely because they are judges.

Courts also recognize that a Facebook friendship is not necessarily a meaningful gesture, and even if it was, judges have friends too and should be able to socialize online (and express social preferences) in the same way they do off-line. (On the legal significance of Facebook friendship generally, see Quigley Corp. v. Karkus, No. 09-1725, 2009 U.S. Dist. LEXIS 41296, at *16, n.3 (E.D. Pa. May 19, 2009): "[T]he Court assigns no significance to the Facebook "friends" reference .. . . Indeed, "friendships" on Facebook may be as fleeting as the flick of a delete button.") This is not to say that judicial officers socializing online do not have to exercise extra care. Privacy settings can be confusing. When viewed offline or in a different context, it's often unclear who online statements may be attributed to. There's also the issue of judicial elections, which present additional concerns.

The real issue here is the ex parte communication, which the judge in this case recognized was a no-no. Most people would know to not send a letter trying to chambers trying to influence the outcome of a case. Does the Facebook system encourage ex parte communications and allow them to take place in a scenario where people who communicate to judges don't see them as obviously problematic? Also, most people would not necessarily have an easy time tracking down a judge's contact information, but does being Facebook friends with a judge make it that much easier to send him or her a note? I don't know the answers to these questions.

In the meantime, litigants will continue to attack decisions based on online relationships, but their efforts are likely for naught.

[image credit: Shutterstock/virinaflora: "best friends"]

UK's New Defamation Law May Accelerate the Death of Anonymous User-Generated Content Internationally (Forbes Cross-Post)

By Eric Goldman

Historically, United Kingdom defamation law has been victim-favorable. In an effort to modernize its defamation law, the UK Parliament recently enacted the Defamation Act 2013 (royal assent was given on April 25). The act generally makes it harder for plaintiffs to win defamation lawsuits, but I'll focus on the effects of Section 5 of the act, entitled "Operators of websites."

Section 5 sets up a "notice-and-takedown" system for defamation from user-generated content (UGC). A web host or other online service provider isn't liable for defamatory UGC unless it receives a takedown notice (containing specific information required by the act) and fails to remove the content on a timely basis. Moderating user content doesn't disqualify the website operator from the act's protection, but the act excludes when the website operator acts with "malice"--a proviso that might prove nettlesome in practice. All told, in theory, a defamation victim cannot sue a UK-based UGC website without first sending a takedown notice and giving the website the opportunity to remove the content.

Notice-and-takedown schemes create predictable problems. It's easy enough for anyone to send a takedown notice, especially because the UK act apparently doesn't impose any adverse consequence for sending bogus notices; and if a website faces liability only if it ignores the notice, it will reflexively remove content in response to takedown notices, irrespective of the notice's legitimacy. So surely the act's notice-and-takedown will be abused, as inevitably happens with all notice-and-takedown systems.

However, what's really interesting about Section 5 is its bonus requirement for UGC websites to avoid defamation liability: they qualify for the act's protection only if the defamation victim can find the user to sue him/her. The act doesn't explicitly say what information about its users the website operator must give a defamation victim or when (and some of these requirements will be spelled out in regulations that are being developed), but to me the implication seems clear: if the website operator can't provide authenticated identifying information about its users, the website operator will lose the act's protection (unauthenticated information is useless to plaintiffs if falsified).

The requirement that website operators help plaintiffs find their users represents a major new development in the law applicable to UGC. In particular, I think the rest of the act's notice-and-takedown scheme effectively codifies existing UK law. For example, a February ruling (Tamiz v. Google, Inc., [2013] EWCA Civ 68) basically adopted a notice-and-takedown rule for defamatory UGC on Google's Blogger.com platform; and the European Union's Electronic Commerce Directive (2000/31/EC, Article 14--apparently adopted by the UK) establishes a notice-and-takedown scheme for all types of UGC, not just defamatory content. So I believe the act's most change to existing law is creating a user-identification obligation.

(Note 1: I don't know EU law well enough to understand if UK imposing this additional requirement violates the EU E-Commerce Directive and, if so, what consequences that has. If you have thoughts about that, please let me know).

(Note 2: some UGC websites, including Facebook and Google+, have voluntarily adopted policies that require users to provide their real names. This law likely won't change their behavior, though it might make those sites feel like they don't have discretion to repeal their policies).

So while we expected the UK Defamation Act to provide additional legal comfort to website operators, instead it creates an apparent dilemma for them. If UGC websites want the legal protection of a notice-and-takedown system, they have to authenticate and identify their users. If they don't, then they lose the act's protection, exposing them to potential liability for defamatory UGC even if they never receive a takedown notice.

Faced with this dilemma, UGC website operators in UK surely will prevent users from publicly posting any content until the website operator has authenticated the user's identity and contact information. (For example, simply having a user's IP address doesn't seem to satisfy the act, nor will websites rely on the fact that it's hard to publish truly anonymous UGC). As a result, even if a user publishes online content "anonymously" or "pseudonymously," their words still will be attributable to them if legally challenged.

Although the act only applies to defamatory UGC, website operators won't authenticate users only when users are writing potentially defamatory content. Instead, website operators will collect and authenticate users' identifying information universally, meaning that data will be available for anyone who submits a qualifying subpoena--including governments and plaintiffs pursuing non-defamation claims. The act could have restricted third party access to this user data when it isn't being sought by a defamation victim, but it didn't.

Further, I expect more legislatures throughout the world will follow this template: conditioning website operator protection (whether based on notice-and-takedown or some other system) on the ability to getting identifiable information about the bad user. So not only will UK websites authenticate all users before allowing them to post UGC, but eventually UGC websites in other countries will do so as well.

All of this reinforces how the United States' Internet Law differs from the rest of the world. The United States does have a notice-and-takedown scheme for copyright infringing UGC (17 U.S.C. 512(c), a provision that's proven somewhat problematic), but the safe harbor doesn't require websites to authenticate users or attribute their content. (Per 17 U.S.C. 512(h), websites must turn over information about their users on request, but they aren't required to collect or keep information about their users). Otherwise, websites generally aren't liable for UGC, whether or not potential plaintiffs can find the users to sue (see 47 U.S.C. 230). Occasionally we've seen proposals in the United States to link website protection with user attribution (see, e.g., New York and Illinois), but (1) those proposals have gone nowhere, (2) they are probably unconstitutional, and (3) when done at the state level, they are preempted by Section 230 (see, e.g., this ruling involving Backpage).

Thus, while anonymous online content appears imperiled globally, websites in the United States should be able to preserve anonymous online content for the foreseeable future. I'll leave it up to you to figure out which approach you like better.

[Photo credit: The Houses of Parliament and Big Ben at night // ShutterStock]

Perfect 10 Gets a Surprising Partial Sumary Judgment in 512 Case--Perfect 10 v. Yandex

By Eric Goldman

Perfect 10, Inc. v. Yandex N.V., 2013 WL 1899851 (N.D.Cal. May 7, 2013)

Few names strike as much fear--and derision--among Internet lawyers as Perfect 10. Perhaps the quintessential Internet Law plaintiff of the 2000s decade, Perfect 10 went on a dubious litigation campaign filled with precedent-setting losses and massive legal bills for all participants. Perfect 10 won a ruling here and there, but for the most part, Perfect 10's courtroom efforts have been a categorical failure. Indeed, few plaintiffs have created more favorable Internet Law precedent than Perfect 10.

Thus, it's a little disorienting/surprising to see Perfect 10 win anything in court, even if it's a small preliminary ruling. Perfect 10's target in this lawsuit is Yandex, the Dutch company operating the leading Russian search engine. In this ruling, the parties are wrangling over Yandex's ability to qualify for the 512 safe harbors.

Consistent with its past (and IMO unnecessarily sloppy) practices, Perfect 10 didn't send a prototypical 512(c)(3) notice. Instead, one takedown notice it sent consisted of:

a cover email attaching the DMCA notice. Next, there was an attached PDF file. The PDF file began with a short DMCA notice letter that described the contents of the DMCA notice itself and requested that the images be taken down. The notice letter was followed by several pages of screen shots from Yandex's own image search web sites. The screen shots showed the allegedly-copyrighted images in Yandex's search results along with corresponding links to the party directly hosting the content. In many instances the links to the third-party sites were truncated, but it was possible to copy the whole link by right-clicking on the image in the file. At the end of each sample notice was a single screen shot from Perfect 10's own website that included the allegedly copyrighted images in a four-by-four grid of images.

Yandex raised several objections to this notice. Yandex says that Perfect 10's notices should be reviewed in combination with the others because the aggregated workload processing the many PDFs it received was unnecessarily high, but the court accepts Perfect 10's request just to look at this one. (That turned out to be an uncharacteristically savvy litigation move from Perfect 10). The court is unmoved by Yandex's beef that it's burdensome to extract the full URLs from the PDF and to match the screen shots to the thumbnails. The court thinks these obligations are within the scope of the service providers' responsibilities, and because Perfect 10 asked the judge to evaluate just one takedown notice in isolation, the court denigrates the amount of work required for Yandex to review each notice in isolation (calling Yandex's gripe "disingenous" at one point). Thus, the court grants Perfect 10's summary judgment request that this PDF takedown notice satisfied 512(c)(3).

(An aside: I find it incredible that Perfect 10's PDF approach was the most efficient way to prepare 512(c)(3) notices. It seems almost like Perfect 10 was trying to satisfy the minimum 512(c)(3) requirements using the most onerous and expensive approach for recipients. They wouldn't be that devious...would they? At minimum, Perfect 10 could have saved itself a lot of future litigation time and expense by sending more traditional notices.)

Yandex also categorically loses any 512(c) safe harbors for the period of time it didn't have a designated agent for service of notice, a step that didn't happen before 2012. Designating an agent is a clear minimum requirement for 512(c) protection, but it highlights the challenges faced by some foreign defendants. It sort of makes sense that a European search engine wouldn't satisfy a US formality, but Yandex was big enough that it should have checked off this box. Whoops.

If history has any predictive power, Perfect 10 will ultimately lose this lawsuit. Still, Perfect 10 got a couple of favorable points out of this ruling. To me, that reinforces the structural design failings of 512's safe harbor.

See more blog posts on Perfect 10.

[Photo credit: Play on ten. Success metaphor // ShutterStock]

Copyright Trolling Is Really Hard to Do Profitably--Righthaven v. Hoehn

By Eric Goldman

Righthaven LLC v. Hoehn, 2013 WL 1908876 (9th Cir. May 9, 2013)

It's been a rough week for copyright trolls. First, Judge Wright destroyed the Prenda Law enterprise in a sanction-filled opinion. Second, today the Ninth Circuit emphatically rejected Righthaven's attempt to manufacture copyright standing. I don't see how either the Prenda Law outfit or Righthaven survive these blows. Both offer a good cautionary tale for anyone who thinks there's gold in copyright trolling. Not only isn't it a path to riches, but it's likely a road to ruin; I'm still waiting to see how many lawyers lose their license from the ill-fated efforts.

The Ninth Circuit says that Righthaven failed to obtain bona fide copyright transfers to Stephens Media's copyrights, despite several attempts to do so. The funny thing is, it would have been possible to satisfy the standing requirement, but only if Stephens Media had been willing to give up control of its copyrights. Instead, Stephens Media simultaneously wanted the cash from Righthaven's litigation campaign but was never comfortable enough actually giving Righthaven any real rights. Using too many double negatives, the court says:

Moreover, the contract evinced not just an intent that Righthaven receive whatever rights were necessary for it to sue, but also an intent that Stephens Media retained complete control over all exclusive rights. The problem is not that the district court did not read the contract in accordance with the parties’ intent; the problem is that what the parties intended was invalid under the Copyright Act.

The final version of the Stephens Media-Righthaven deal (contained in a “Clarification and Amendment to Strategic Alliance Agreement") said that Stephens Media "convey[ed] all ownership rights in and to any identified Work to Righthaven through a Copyright Assignment so that Righthaven would be the rightful owner of the identified Work.” Yet, even this contract provision failed because Righthaven had to give Stephens Media 30 days notice before exploiting the work, and Stephens Media could buy the rights back by giving 14 days' notice and paying $10. The court says:

Consequently, Righthaven was still unable to exploit any exclusive rights unless Stephens Media permitted it to. Meanwhile, Stephens Media was free to exploit the works to the full extent it wished, and it presumably would with any article that it perceived to have additional value.

Sounds like Righthaven could have benefitted from a real copyright expert to handle these provisions. Where was Dale Cendali when Righthaven needed her most?

Because Righthaven never properly procured a copyright transfer to any of Stephens Media's works, it lacked standing for all of its lawsuits based on Stephens Media's works. In turn, because Righthaven lacked standing, the Ninth Circuit vacates the lower court fair use rulings in the Hoehn and DiBiase cases (and presumably Righthaven's other fair use losses are similarly suspect), an unfortunate but logical collateral consequence.

In theory, Righthaven could try once again to acquire proper copyright transfers from Stephens Media and refile all of its lawsuits, but that won't happen. It's clear Righthaven didn't know how to litigate properly, it's clear that there's not a lot of easy profits in suing mom-and-pop bloggers who may unintentionally violate a newspaper's copyright, and Righthaven has lost all of its assets. So Righthaven is dead, and it won't be resurrected. Good riddance. The remaining mop-up issues from the Righthaven debacle include whether Stephens Media will have to cough up more money for its role in the sordid affair, and how the various state bar discipline enforcers view the conduct of some of Righthaven's attorneys.

In September 2010, I had a telephone "debate" with Steve Gibson of Righthaven. In the call, I kept pounding the point that I didn't believe his scheme could be profitable. Over the last three years, I haven't seen anything that changes my view. Copyright trolling requires a high volume of low-value cases, which requires lawyers to keep costs down--inevitably by cutting legal corners. Thus, there is an intrinsic tension between keeping costs down and generating enough bona fide copyright infringement cases to get paid off. Good data points for future copyright owners who thinks trolling is the path to riches.

Prior blog posts:

* Righthaven Hit With Another Fee/Cost Award, This Time Nearly $120k--Righthaven v. DiBiase
* Colorado Judge Drills Righthaven and Awards Attorneys' Fees--Righthaven v. Wolf
* Resetting the Righthaven Fiasco
* Righthaven Defendant Awarded $3,800 in Attorneys' Fees--Righthaven v. Leon
* Recapping Righthaven Developments from the Past Two Weeks
* Righthaven Benchslapped in Ruling Saying It Lacks Standing--Righthaven v. Democratic Underground
* Another Defense-Favorable Righthaven Ruling--Righthaven v. Choudhry
* Republishing Entire Newspaper Story is Fair Use--Righthaven v. CIO
* Blogger Wins Fair Use Defense...On a Motion to Dismiss!--Righthaven v. Realty One

[Photo Credit: trolling road sign // ShutterStock]

Yahoo's User Agreement Fails in Battle Over Dead User's Email Account--Ajemian v. Yahoo

[Post by Venkat Balasubramani]

Ajemian v. Yahoo!, 12-P-178 (Mass. Ct. App. May 7, 2013)

This is a very interesting dispute that raises the question of ownership over digital assets after a person’s death.

Plaintiffs, John Ajemian's (the decedent's) executors and siblings, sued to declare his estate owner of email messages he sent and received via his Yahoo! account. Prior to filing suit, they tried to negotiate with Yahoo! to get access to the account, among other things, to notify friends of John’s death and service arrangements, and to organize and administer John’s assets. Yahoo! co-operated in providing basic subscriber information as part of an initial lawsuit, but Yahoo took the position that the Stored Communications Act barred disclosure of the message contents. A probate judge dismissed the complaint, finding that the forum selection clause required the lawsuit to be brought in California. The appeals court reverses.

Were the contract provisions reasonably communicated to the end user: On the core question of whether the online agreement is enforceable, the court says that Yahoo!’s terms should be enforced under the standards of any other agreement. The key question is whether the terms had been reasonably communicated to the end user. The terms had been amended to add a no third-party beneficiary and a “no-survivorship” clause. The court says the record is unclear as to how the initial terms, and the revised terms, were communicated to users. Because of the lack of clarity in the record on this issue, Yahoo! is not entitled to dismissal based on the forum selection clause, or based on other terms that would limit the estate's rights vis-a-vis the account.

Is the forum selection clause enforceable against the administrators?: Even assuming the terms and revisions were adequately communicated to the decedent, the court says that it would be unreasonable to enforce the forum selection clause against his estate administrators. The administrators are not parties to the agreement, which references the singular “you” in describing the counter-party to Yahoo! The decedent was domiciled in Massachusetts, and probate courts in the State of Massachusetts have a strong interest in resolving ownership questions over a resident decedent’s assets. Because the lawsuit involves a narrow issue about the account contents of the account, Yahoo!’s interest in having the lawsuit resolved in California is not as strong as it otherwise might have been (e.g., if the lawsuit was about Yahoo!’s services). The court also says that the forum selection clause is overly broad, and as written requires any suit between a subscriber and Yahoo! to be brought in California.

(The probate judge concluded that Res Judicata barred the second complaint which sought the contents of the email; the appeals court reverses on this question as well.)

An interesting factual backdrop to the case is that as the court notes, one of the administrators actually helped the decedent set up the account and may have even “shared” the account with the decedent. (Unfortunately, he did not remember or have the password.)

The merits are a veritable thicket. Does the Stored Communications Act allow an estate to grant consent on behalf of the decedent? Earlier cases involving Facebook address the waiver issue, but none really resolve it. (Stored Communications Act Bars Disclosure of Facebook Records to Surviving Family Members in the UK; "Court Orders Facebooking Juror to Disclose Additional Facebook Posts--Juror No. 1 v. Superior Court".) There's also the distinction between copyrights in the communications and ownership of the account (i.e., the "chattel"). While the estate probably owns the copyrights to the account contents, this does not mean that it can force Yahoo! to provide access to the account itself.

It's worth looking at what (if anything) Yahoo! could have done differently here.

1. Better contract amendment process: It's tough to say definitively whether the terms of service may be vindicated down the road, but the court’s approach to the terms of service issue reflects a fair amount of skepticism towards online agreements. While the court pays lip service to the fact that online agreements should be treated the same as any other contract, the court engages in some judicial contortion to not enforce the contractual terms. I'm not sure Yahoo! could do much more on this front to change the result here. On the other hand, Yahoo!’s terms contain the unfortunate “we can amend this agreement without providing you notice” language that companies would be wise to avoid.

2. Interplead the contents: Could Yahoo! have deposited the emails in the court’s registry and just have abided by the court’s decision? I'm not sure the Stored Communications Act envisions this, but it's not a great alternative for Yahoo! anyway. On a long term basis, this would mean that it will have some administrative involvement in far-flung jurisdictions.

3. Create a mechanism to allow users to control the fate of their accounts post-death: Google recently offered an “Inactive Account Manager,” that lets people designate what happens to the accounts when they pass. (See Kash Hill’s post: “Will You Use Google's Death Manager To Let Loved Ones Read Your Email When You Die?”.) This sounds like a good solution, although it requires an investment of resources on Yahoo!'s part.

Yahoo! could have, per its terms of service, merely deleted the content altogether, but it took a relatively consumer-friendly approach and preserved the contents. Unfortunately, as a result, it's now embroiled in an ongoing dispute in probate court in Massachusetts.

[image credit: Shutterstock / JMiks - "Login Box"]

Suing Like It's 2009: Parts.com Sues Google and Yahoo for Keyword Advertising

By Eric Goldman

Parts.com v. Google, 3:13-cv-01074-JLS-WMC (S.D. Cal. complaint filed May 6, 2013); and
Parts.com v. Yahoo, 3:13-cv-01078-AJB-JMA (S.D. Cal. complaint filed May 6, 2013)
_____

In the immediate wake of the Second Circuit's Rescuecom ruling in 2009, about a dozen trademark owners sued Google for selling keyword ads. Google resolved each and every one of those cases, and a couple more since then, without suffering a final adverse ruling in any of them. Following its important settlement with Rosetta Stone, I believe Google now has only three pending trademark lawsuits over keyword advertising, all of which seemed doomed to me: the CYBERsitter and Home Decor Center cases, both filed in 2012 in the wake of the Fourth Circuit's Rosetta Stone ruling, and the quixotic Carla Ison case in state court.

Earlier this week, Parts.com joined this dubious plaintiff-vs.-Google club (and it sued Yahoo for good measure). There's nothing fancy or creative about Parts.com's lawsuit--it's a direct, and probably futile, frontal assault on Google's AdWords cash cow that generates around $40B a year. Some of the interesting angles to this lawsuit:

* Parts.com is a terrible trademark, about as protectable as Pets.com (which trademark lawyers often use as a prime example of a completely untrademarkable phrase). Indeed, there's good reason to believe Parts.com is generic following the Advertising.com, Hotels.com and Mattress.com rulings. If Parts.com doesn't voluntarily drop this lawsuit (a surprisingly common outcome for AdWords challengers), I predict Parts.com ends this lawsuit with an invalidated trademark; similar to how American Blinds ended its Google lawsuit with fewer trademarks. I really don't understand the business rationale for investing dollars in litigation where a likely outcome is having a judge conclude that your assets are worthless, but more than one trademark owner has asked Google to help it achieve this dubious outcome.

This lawsuit fits a trend that I've mentioned before: trademark owners with crappy trademarks are often unusually pugnacious about enforcing their purported trademark rights. 1-800 Contacts is my premier example of that phenomenon. Some of the "trademark" owners that have challenged Google include Home Decor Center, American Blinds, and now Parts.com. See the trend?

* Parts.com says it first contacted Google in 2007 to complain about its keyword ad sales. Trademark doesn't have a statute of limitations, but six years is a mighty long time for this issue to fester, so I could see a strong laches defense. See, e.g., the Southern Grouts v. 3M case, where laches barred a five year delay in enforcement.

Especially perplexing is that Parts.com got its trademark registration in Sept. 2008, so if it waited four more months, its trademark should have become incontestable. Incontestability wouldn't overcome the serious genericism problem it faces; but if you're going to wait six years to bring a suit, why not wait a few more months to get the limited benefits of incontestability....before poking at giants who will deploy massive resources to destroy the trademark?

[I emailed Parts.com's counsel about both of these issues but didn't get a response.]

* I am unclear if Google broad-matches the noun in a "[noun].com" query. So, in this case, does Google deliver ad results based on just the keyword "parts" if someone does a search for "parts.com"? I've asked a few folks but haven't got a reliable answer yet. If Google broad-matches this way, then Parts.com has even less to complain about. (I don't know Yahoo's broad-matching policy, but I'd welcome input on that too).

One final oddity: this is the second time that Google has been sued by an online parts retailer for trademark infringement due to keyword advertising (PartsGeek was the first). Selling parts online must be a pretty ugly business if suing Google is a good business decision. Parts.com claims it can prove losses of $2M due to infringing keyword advertising, but I'm calling BS on that. Based on the numbers we've seen in other cases, my guess is that the total value of "diverted" clicks--if any exist at all--is more properly measured by pennies, not millions.

____

Pending roster of AdWords trademark lawsuits against Google:

* Parts.com v. Google
* CYBERsitter v. Google [Latest post]
* Home Decor Center v. Google
* Carla Ison v. Google [the appellate docket]

____

My recent Tertium Quid blog posts on keyword advertising lawsuits:

* Florida Proposes to Ban Competitive Keyword Advertising by Lawyers
* More Confirmation That Google Has Won the AdWords Trademark Battles Worldwide
* Google's Search Suggestions Don't Violate Wisconsin Publicity Rights Law
* Amazon's Merchandising of Its Search Results Doesn't Violate Trademark Law
* Buying Keyword Ads on People's Names Doesn't Violate Their Publicity Rights
* With Its Australian Court Victory, Google Moves Closer to Legitimizing Keyword Advertising Globally
* Yet Another Ruling That Competitive Keyword Ad Lawsuits Are Stupid--Louisiana Pacific v. James Hardie
* Another Google AdWords Advertiser Defeats Trademark Infringement Lawsuit
* With Rosetta Stone Settlement, Google Gets Closer to Legitimizing Billions of AdWords Revenue
* Google Defeats Trademark Challenge to Its AdWords Service
* Newly Released Consumer Survey Indicates that Legal Concerns About Competitive Keyword Advertising Are Overblown

[Photo credit: New year's odometer // ShutterStock]

Crazy SOPA-Like Attempt to Hold International Banks Liable for Pharmacy Spam Fails on Jurisdiction Grounds--Unspam v. Chernuk

[By Venkat Balasubramani with comments from Eric Goldman]

Unspam Technologies, Inc. v. Chernuk, 2013 WL 1849080 (4th Cir. May 4, 2013)

We’ve mentioned “Project Honeypot,” the efforts of a company (founded by Matthew Prince) to track down and prosecute spammers. This lawsuit was ambitious in its scope and suitably quixotic. Like SOPA, it sought to attack pharmacy spam by going after the money.

Unspam (operator of Project Honeypot) and John Doe sued two individuals for pharmacy spam. They also named a slew of international banks who allegedly provided transaction processing services to pharmacy spammers. The banks submitted affidavits demonstrating that they had no activity or customers in the United States, and therefore they should not be subject to personal jurisdiction in the U.S. The district court agreed, granting the banks’ motion to dismiss on jurisdictional grounds. Unspam appealed to the Fourth Circuit.

Unspam fared no better on appeal. In a published opinion, the Fourth Circuit affirms the district court’s dismissal of the banks on jurisdictional grounds:

Not one of the banks directed its business to Virginia or aimed its commercial efforts at customers in Virginia. Indeed, there is no evidence that any drug transactions involving the plaintiffs were connected by intermediaries to these banks. Moreover, even if we were to assume that Doe's purchase was presented by some Internet “pharmacist” to one of the foreign banks for processing through the international Visa network, that transaction still would be too remote an act to justify jurisdiction in Virginia. The transaction would have occurred in the foreign country where the pharmacist presented the Visa charge to the bank, and thereafter, the bank would simply have collected the charge through the Visa network. The foreign bank's relevant activity would thus be localized to the foreign country where it did business, and its only conduct “aimed” from that location would be the transmittal of the transaction into the Visa network. The fact that the transaction ultimately rippled through other countries for the collection of monies would not indicate that the bank purposefully availed itself of the laws of the countries where subsequent transactions occurred..

Plaintiffs also argued that they need not establish jurisdictional facts over the banks because the banks were part of a conspiracy and were subject to jurisdiction based on the acts of co-conspirators. Plaintiffs argued, relying on “blog research and internet searches” that “Canadian Pharmacy” was a trade name for two of the defendants and that Chronopay [described as “the PayPal of Russia”] provided transaction processing services to defendants, with the defendant banks acting as links in the chain. The court says this is mere speculation, and there is no evidence or allegation regarding the transaction(s) involving the doe plaintiff.

Finally, plaintiffs tried to rely on Rule 4(k), which provides personal jurisdiction based on federal claims where the defendant may not be subject to jurisdiction in a particular state, but has minimum contacts with the U.S. as a whole. The court says that defendants do not have sufficient contacts with the United States as a whole for plaintiffs to invoke Rule 4(k) as a basis of jurisdiction.
_____

As mentioned in my post on the district court ruling (Is SOPA's "Follow the Money" Meme Infecting Anti-Spam Litigation? – Project Honey Pot v. Does), on the merits, plaintiffs' claims were a serious stretch. It’s not surprising that the plaintiffs got zero sympathy from the Fourth Circuit on the jurisdictional issues.
_____

Eric's Comments. The Fourth Circuit reached the logical result in this case. However, I view this lawsuit as a cautionary tale of how close we are to breaking the Internet.

The overzealous war against Internet pharmaceuticals has already taken a half-billion dollar toll on Google and another $40M toll on UPS (yes, the shipping company). In both cases, these intermediaries "agreed" (with the DOJ's axe ready to swing if they didn't acquiesce) to become the DOJ's deputies and police their customers more aggressively. The result is that we now have key cogs in the wheel making legal determinations about the legitimacy of their customers' activities, with zero due process or other oversight of any errors they make make against customers.

This problem only exacerbates as we continually expand the list of potential deputies who should be combating illicit activity online. There is no natural boundary for that witchhunt, and lots of defendants got tossed into the litigation mix and have to spend lots of money defending their conduct (even if legitimate). Worse, any vendor who might be deputized becomes adversarial to their customer base, with the inevitable waste of resources and chilled environment for innovation. For those of you who think that a narrow "follow the money" liability trail as an alternative to more draconian intermediary deputization, I encourage you to review this lawsuit and think carefully if this is the world you want. For more on this, see my six month retrospective of SOPA.
_____

Previous post:

Is SOPA's "Follow the Money" Meme Infecting Anti-Spam Litigation? – Project Honey Pot v. Does

[image credit: Shutterstock - Matthew Cole - "Illustration of a Honey Bee on a White Background"]

Assessing Winners and Losers in Google's Worldwide Antitrust Battles (Forbes Cross-Post)

By Eric Goldman

Recently, the European Commission announced details of a proposed antitrust settlement with Google (the full commitments). Google's competitors now have the opportunity to publicly whine about the deal as part of the "market testing" procedural step. I doubt the inevitable sour-grapes objections will prompt any meaningful change to the settlement.

The EC's proposed antitrust resolution with Google follows the FTC's resolution of a similar inquiry earlier this year. With these matters nearing the finish line on both sides of the Atlantic, I thought it might be timely to assess winners and losers from Google's multi-year antitrust battles.

Winner: Google. Poor Google. They have suffered through one low-merit antitrust legal challenge after another, including federal and state inquiries and private litigation (the latter has gone nowhere). Of course, competitor (and regulator) jealousy and greed come with the territory any time a company finds a way to mint money. But don't feel too sorry for Google; many of us would happily endure its antitrust problems for even a small sliver of Google's annual profits.

Despite its wasted time and money, viewed holistically, Google has emerged as a big winner of its multi-year, multi-front antitrust battles. Its competitors and critics have repeatedly hit it with their best shots, yet Google's aggregate concessions have been minimal. Google's antitrust legal team--both employees and outside lawyers and consultants--have done a remarkable job quelling the threats they've faced.

Winner (Mostly): FTC. The FTC made two choices that should have chalked it up as a loser. The FTC never had the evidence to support a broad antitrust inquiry into Google's search practices, and the inquiry lasted longer than it should have. Yet, I'm counting the FTC as a winner for two reasons. First, it decisively admitted that it never found adequate evidence of Google's search bias, rather than trying to keep bluffing a weak hand. Second, the concessions it did obtain from Google were generally narrowly tailored for the problems the FTC sought to redress.

Loser: European Commission. Superficially, the EC had more antitrust leverage against Google than the FTC. EU antitrust law is more pro-regulator than US law, and Google has a much greater share of the search market in Europe than in the US. Yet, despite years of tough rhetoric from EU antitrust regulators, Google's concessions to the EC are only a little more than what the FTC got. Viewed that way, I'm tempted to say the EC "underperformed" compared to the FTC. However, given that neither the EC nor the FTC ever had any substantial grounds for their antitrust inquiries, the EC's real failing was publicly promising more than it ever could deliver.

Loser: Google's Competitors and Critics, including Microsoft, the questionably named "FairSearch," vertical search engines and online content publishers. For years, I've felt like Google's competitors waged a war of attrition against Google, trying to tie down Google in as many sticky battlegrounds as possible. If that were the only goal, the competitors' campaign has been successful. However, overall, a war-of-attrition against Google never made sense. Google's financial fortunes allow it to easily outlast the most determined rivals. Further, if the goal was to hamper Google's innovation process, that also failed decisively. Not only has Google continued to evolve and improve its core online services, but Google continues to generate lots of buzz for things like Google Glasses and driverless cars.

So what do Google's competitors have to show for their multi-year campaign against Google, and the (tens of?) millions of dollars spent along the way? A few restrictions on Google's practices, mostly de minimis to Google and worthless to its competitors. The most significant "payoff" has been the regulatory mandate that Google create and display a new ad unit alongside its own vertical search results. The ad unit will provide links to up to three competitive vertical search services that meet minimum standards; in some cases, the links will be free, otherwise the links will be auctioned among qualified search services. Exactly how valuable is this new ad unit to advertisers/Google rivals? While advertisers always love free traffic, my guess is that consumers will largely ignore these links and they will deliver minimal traffic, in which case there won't be much bidding at the auctions. I hope we'll see data about the clickthroughs and auction prices eventually so we can judge just how useful this concession really was.

Loser: a Unified Internet. Google will implement some of its EC commitments only in the European Economic Area (the EEA). Thus, EEA consumers will experience a different Google than the rest of the world (though the differences aren't major). For Internet purists who once hoped that the Internet would help integrate a unified global community, the additional regulatory-driven deviations between the European Google and the rest-of-the-world Google has to be disheartening. Indeed, due to different regulatory schemes, many global Internet properties like Google already run country-specific flavors of their services, reifying existing geographic differences rather than transcending them.

Winners: Consumers (But Not For the Reason You Think). The EC commitments mandate that Google follow certain user interface designs, such as displaying a box around certain Google self-promotional links and displaying icons next to certain search results. The underlying subtext is that such efforts will improve consumer search processes, but is this true? Typically Google does A/B user testing before changing any user interfaces to see if new interfaces are an improvement over the existing ones. Was A/B testing done here? Sadly, nothing indicates that any consumer testing was done, and I suspect the EC guessed what consumers wanted rather than basing their mandates on actual empirical data. If I'm right, it's possible the EC's user interface mandates will backfire. For example, one possibility is that the required boxes and icons will draw users' eyes to the Google links, which will counterproductively drive more internal traffic to Google at the expense of its competitors.

Either way, I don't expect anything in the commitments (or, for that matter, the FTC settlement) will make a material difference to consumers. If I'm right, these resolutions don't harm or improve consumer welfare; they will simply reflect a lot of regulatory angst to achieve the same outcome. Based on my expected results, I'm declaring the global Google antitrust battle resolutions to be a non-event for consumers. Given the risk that regulators were going to screw up something as important as online search, that's actually an unexpected win for consumers.

More Reading: I've written two articles about antitrust regulation of Google's search results: Search Engine Bias and the Demise of Search Engine Utopianism from 2006 and Revisiting Search Engine Bias from 2011.

[Photo Credit: Bunch of moldy red grapes // ShutterStock]

Recap of Washington State’s Employer Social Media Password Bill

[Post by Venkat Balasubramani]

SB 5211 (passed by Senate April 27, 2013)

Both chambers of the Washington legislature passed Substitute Senate Bill 5211, and it now awaits the Governor’s signature. No Washington legislators voted against it.

Here are the key provisions:

- the bill restricts employers from (1) asking for passwords; (2) engaging in shoulder-surfing; (3) asking employees to change settings; or (4) asking employees to add them as friends;

- the statute does not define it, but uses the phrase “personal social networking account” to describe what is off limits;

- the statute contains a carve-out for employer investigations, but limits this to scenarios where employers merely request information (and not passwords) in order to investigate (1) legal violations or (2) suspicion of misappropriation by employees;

- the statute excludes employer-supplied devices, as well as intranets and other platforms “that [are] intended primarily to facilitate work-related information exchange, collaboration, or communication by employees or other workers” [Twitter!];

- finally, the statute provides for a private right of action and authorizes a court to award $500 in statutory damages and attorney’s fees.

The statute tackles the social media ownership question by excluding accounts that are “provided by virtue of the employee’s employment relationship with the employer .. [or] online account[s] paid for or supplied by the employer.” Earlier versions of the statute had more specific definitions for what constitutes a social network. One iteration also limited the private right of action.

Although I acknowledge that employee social networking activity that is private should be off-limits for employers, I still think this was a solution in search of a problem. With states’ well established failings in the era of regulating the internet, I predict this will create more problems than it will solve. Among other things, it's unclear what is covered by "social media" and equally unclear when an account is "personal" or "business"-related. (In practice they often seem mixed.) Eric delves into some of these issues in taking a look at California's social media legislation in this post: "Big Problems in California's New Law Restricting Employers' Access to Employees' Online Accounts."

Two approaches legislatures should consider taking when enacting these statutes is (1) making them only apply to prospective employees; and (2) getting rid of a private right of action or, at least, including an administrative exhaustion component. The Littler law firm has a great roundup of legislation from various states as well as some of the pitfalls these statutes may present: "Social Media Password Protection and Privacy -- The Patchwork of State Laws and How it Affects Employers" [pdf].

Other coverage on employer social media bills generally:

Littler: "Social Media Password Protection and Privacy -- The Patchwork of State Laws and How it Affects Employers" [pdf]
WIlliam Carleton: "Checking in on sate social media password laws"
Eric Meyer: "Wooooo pig sooie! Arkansas gets a workplace social media privacy law"
Lexology "Three more states hop on the social media legislation bandwagon"

[image credit: shutterstock/Andre Blais "high-tech background with targeted eye-scan"]

More Evidence That Congress Misaligned the DMCA Online Copyright Safe Harbors--UMG v. Grooveshark (Forbes Cross-Post)

By Eric Goldman

UMG Recordings, Inc. v. Escape Media Group, Inc., 2013 WL 1729431 (N.Y. App. Div. April 23, 2013)

Grooveshark runs a user-generated content (UGC) website that allows users to upload sound recordings and other users to stream those recordings. UMG Recordings (a subsidiary of Vivendi, VIV:FP) sued Grooveshark for copyright infringement based on its users' activities. This week, a New York state appellate court partially denied Grooveshark's eligibility for the online safe harbors enacted by Congress in the 1998 Digital Millennium Copyright Act. This ruling creates several problems for UGC websites that let users post sound recordings, and it demonstrates one way Congress mishandled drafting its online safe harbors.

The Ruling

In 1998, Congress provided UGC websites with a safe harbor for user-caused copyright infringement (17 U.S.C. 512(c)). To be eligible for the safe harbor, UGC websites must satisfy some preconditions. If they do, copyright owners unhappy about users committing copyright infringement can send takedown notices to the UGC website. If the UGC website responds quickly to a takedown notice, the DMCA online safe harbor says that the website avoids further copyright liability for the items identified in the takedown notice. In contrast, if the UGC website ignores/rejects the takedown notice, it potentially bears legal responsibility for its users' actions. Obviously, most UGC websites prefer to limit their risk, so they routinely take down items identified in the takedown notices.

I acknowledge that you probably think most court opinions address arcane legal issues, but even I think this ruling dealt with unusually arcane legal issues. The DMCA online safe harbors apply to "copyright" claims, but it turns out the word "copyright" is ambiguous. In the 1976 Copyright Act, Congress attempted to eliminate most state copyright laws. As a result, today most potentially copyrightable works either are protected under federal law, or they are not protectable under federal or state copyright law at all. However, Congress preserved a few categories of works that can be covered by state copyright law--the most commonly-encountered example is certain bootleg recordings of concerts. Congress also said that sound recordings made before 1972 remain protected under state copyright laws. Some of these pre-1972 sound recordings are hugely important and highly recognizable cultural assets; for example, the Grooveshark court cites the UMG-owned recordings of “Peggy Sue” by Buddy Holly, “Johnny B. Goode” by Chuck Berry, “My Girl” by the Temptations and “Baby Love” by the Supremes.

It's this latter group of pre-1972 sound recordings at issue in the Grooveshark case. Grooveshark argued that the DMCA safe harbors applied to both federal and state copyrighted works. UMG argued that the safe harbors only apply to federal copyrighted works, not state copyrighted works. In a short and relatively unenlightening opinion, the New York state appellate court sided with UMG and ruled that the DMCA safe harbor's "notice-and-takedown" scheme doesn't apply to state copyrighted works.

Implications

I see at least three problems with the court's ruling:

Problem #1: In 20111, a New York federal court (in Capitol v. MP3Tunes) reached the directly opposite conclusion and held that state copyrighted works are covered by the DMCA. (The Grooveshark opinion acknowledged this precedent but made no effort to distinguish it). In 2007, the federal Ninth Circuit Court of Appeals held in 2007 that all state IP claims against UGC websites (including, presumably, state copyright laws) categorically are preempted by a different federal website immunity, 47 U.S.C. 230 (see Perfect 10 v. ccBill).

Thus, we have three different and conflicting interpretations of the DMCA's applicability to pre-1972 sound recordings. What a mess! It's never good for anyone when courts have three different answers to the same legal question. It also means litigants may engage in wasteful forum-shopping efforts to find the judicial venue where the rules are most favorable them.

Problem #2: The ruling doesn't hold Grooveshark liable for users' copyright infringement of pre-1972 sound recordings. Instead, Grooveshark simply failed to qualify for the safe harbor for those works, so the judicial inquiry will now turn to the default laws applicable to "secondary" infringement of state copyrighted works. Unfortunately, we have no idea what those rules are. Because most UGC-related copyright infringement cases have been resolved by the DMCA safe harbor, we don't have many rulings interpreting the secondary infringement rules in those rare situations when the safe harbor doesn't apply (see this post for more about that). Furthermore, we have even scarcer caselaw interpreting secondary liability for infringement of state copyrighted works. (It's probable the rules will track the federal copyright laws, but that's not guaranteed). Thus, this case now involves novel and unpredictable legal questions.

Problem #3: Grooveshark has no easy way to distinguish which user-submitted sound recordings are covered by federal copyright versus state copyright. However, its legal liability depends on this difference. Operationally, how should Grooveshark proceed? Without universal protection from the DMCA's safe harbor "notice-and-takedown" scheme, Grooveshark may be required to pre-screen user uploads to assess whether the file is a pre-1972 sound recording or not, and then it may have to handle those files differently.

However, if that's the result, Grooveshark probably doesn't have a viable business. The pre-screening costs would be exorbitant, Grooveshark would make many classification errors, and copyright owners would subsequently argue (probably unsuccessfully, though we are not sure) that Grooveshark should lose the DMCA safe harbor if it fails to catch infringing files during its pre-screen. (See the latest Viacom v. YouTube ruling for examples of copyright owners' arguments about website operator scienter).

More importantly, if Grooveshark must build a uniform site-wide operational process to deal with the small minority of user-uploaded recordings protected by state copyright law, then the DMCA online safe harbors failed at a fundamental level. Even if the safe harbor applies to 99% of the works, the safe harbor doesn't help UGC websites with their business planning because the UGC website must still anticipate and address the 1%--and this becomes impossible if, without further research, the 1% looks identical to the 99%.

In a recent post, I discussed some key design attributes of safe harbors and immunities. One element I identified was "global preemption," meaning that a successful safe harbor has to swipe out all overlapping claims covering the same activity. The Grooveshark ruling suggests Congress failed that design principle by leaving open a state copyright workaround. In my prior post, I identified some other ways that the DMCA online safe harbors violated best-practice design principles. I doubt Congress will be revisiting the DMCA online safe harbors soon (nor would I anticipate any beneficial changes when it does), but its mistakes with the DMCA safe harbors offer useful lessons for Congress' next attempts to draft effective safe harbors.

For more discussion about the DMCA online safe harbors, see the recordings and resources from the recent 15 Year DMCA Retrospective Conference at Santa Clara University.

Differences Between Consumer Surveys for Trademark Cases and False Advertising Cases

By Eric Goldman

In February, I spoke about the differences between consumer surveys in trademark cases compared with false advertising cases. My talk notes:

Overall, the similarities between consumer survey in the two types of cases outweigh the differences. So some of these differences are necessarily exaggerated.

Structural Differences. Lanham Act false advertising cases always involve competitors. Trademark litigation can involve a broader range of defendants, not just competitors. Two examples of consumer surveys in non-competitive trademark cases: dilution surveys; and establishing secondary meaning in non-competitive cases such as, say, parodies. I don't know that this structural difference translates into a lot of meaningful differences in the surveys, but it changes the cases' complexion and may affect how the litigants approach the surveys.

A second possible structural difference. Consumer surveys are effectively mandatory in implied falsity cases. In contrast, although surveys are important for establishing secondary meaning, many successful trademark cases never include a consumer survey.

Range of Inquiry. In both false advertising and trademark surveys, consumers are surveyed about what meaning they derive from the specified stimuli. However, false advertising has a broader range of possible meanings that a consumer could ascribe. In trademarks, consumers are often asked if they recognize a mark or if the mark is similar to or associated with another mark. These questions are basically binary yes/no questions.

False advertising doctrine is interested in what message(s) consumers received. A single claim could have a multiplicity of meanings to consumers--consumers bring heterogeneous perspectives to the claim, and words and images have numerous possible meanings.

The result is that false advertising survey questions tend to be more open-ended and require multiple iterative questions (progressively becoming more close-ended) to ferret out the consumer's interpretation of the claim.

Furthermore, a single product may appeal to multiple consumer segments, so more expensive tests may be required to distinguish each segment and determine what message(s) they received.

Testing Environment. A false advertising survey typically seeks to determine what decision consumers will make in the marketplace in response to the claims at issue. Thus, every little fact about the testing environment matters. This is true with trademarks as well, but false advertising claims evaluation typically draws on a wider range of potentially relevant facts. In false advertising, there may be multiple versions of ad copy or claims that need to be tested together. The tester wants to replicate the buying environment as closely as possible, though a testing environment typically is stylized. The tester also wants to present controls that differ only with respect to the claims at issue. This can be a challenge; it may be difficult to manufacture an appropriately controlled ad copy or control specimen.

[Photo Credit: brand awareness survey // ShutterStock]