A few more ICE tidbits courtesy of FOIA

As the result of ICE's continued production, I've been able to review more from the sample of incidents we sought.  The key takeaways so far have already been covered--TM owners are the source of guidance, and they overreach, but most merchandise that's seized is not parodic.

You're not even trying, buddy: this counterfeit authenticity claim is so incoherent that one wonders why they bothered.  Accept no substitute, indeed:

Previously redacted NFL claim that parody/disparaging shirts are counterfeit:

Also, the relevant industry organization, CAPS, claims that counterfeit clothes pose health risks, as evidenced by this picture.  I''m not saying this is attractive storage, but it is in plastic, and I take my clothes into the bathroom (not even wrapped in plastic) all the time:

(CAPS also makes the claim on its website, which represents the NBA, NFL, MLB, NHL, and collegiate liceonsors, that derogatory, lewd, or offensive uses are counterfeit.)

This series of before and after pictures indicates some selectivity in seizures, which is good.  I note that my Dr. Seuss Slammed I Am T-shirt (which I had, coincidentally, for a number of years before I brought this case) was not seized:

before

before

before

after

seized

after

Documents from my ICE suit

Available here, via Public Citizen. From what we have, I think it's fair to say that most seizures are legitimate counterfeits, so to speak.  That doesn't excuse ICE or the brand owners who supply them with information from the obligation to comply with the law, of course, especially when the exceptions are particularly likely to serve free speech values.  But it's much better than the alternative.

Another thought on ICE's statements

This isn’t specifically IP-related, but I’m pondering the part of my FOIA suit requesting training materials for ICE agents relying on case law, etc.  Initially, the ICE spokesperson who responded to my inquiry said that ICE based its seizure decisions on consultation with agency and Department of Justice attorneys, including consideration of “potential fair use provisions and federal circuit-specific case law.”  When ICE didn’t produce any documents relating to any of that, we used the initial statement as evidence that ICE hadn’t met its FOIA obligations.  The court found it perfectly plausible that all ICE used to determine counterfeit status were manufacturers’ guides, rather than materials of its own.

Now, I will admit that I too suspected that the initial claims were bullshit, in the specific sense of being made with complete indifference to their truth or falsity in order to achieve a non-truth-related conversational aim, that of getting me to shut up and go away.  (That worked really well, as you can see.)  But it’s depressing—I wish I could say extraordinary—that the court didn’t even count that statement as evidence that such guidance existed.  What does it mean for a democracy when a court tells us we shouldn’t take official statements seriously, even—maybe especially—statements made at such a low level about matters that are relatively low-stakes?  Is government only to be trusted when it's under oath?

NFL is advising ICE to seize obvious parodies, my FOIA suit reveals

When last I reported on this, the judge had ordered ICE to unredact materials showing company advice to ICE to seize parody merchandise, since such merchandise wasn't counterfeit and there could be no legitimate law enforcement interest in keeping this advice from the public.  Indeed, ICE's subsequent production showed that, like the NHL, the NFL has instructed ICE that obvious parodies/nonconfusing uses are counterfeit:

From NFL manual: Dallas Sucks shirt is counterfeit because "NFL licensed merchandise will never favor one Club over another or make derogatory use of another Club's marks"

Another supposedly counterfeit shirt showing Minnesota player urinating on Green Bay Packers logo inside outline of Wisconsin

So, what do we know?  (1) Despite ICE's initial claims to me via its spokesperson and a lawyer, ICE relies only on industry guides to identify counterfeits, not on any independent sources.  (2) Those industry guides identify what they don't like, not what is within the scope of counterfeit goods.  (3) Most of what ICE seizes is truly counterfeit, but when it seizes parody merchandise, it implicates First Amendment interests in free speech.

What next?  (1) ICE should improve its guidance to agents, though I'm not sure how that is best to be accomplished.  (2) Industry guides should limit themselves to true counterfeits.  I will note that thus far it doesn't seem that clothing brands, which also provided guides to ICE, overreached in the same way, so it can be done.  Calling the NFL/NHL!

Law360 article on my ICE suit by Bill Donahue

Read it here.

Opinion in my suit against ICE

Tushnet v. U.S. Immig. & Customs Enforcement, No. 1:15-cv-00907 (D.D.C. filed Mar. 31, 2017)

Thanks to my able counsel Michael Kirkpatrick of Public Citizen, initially assisted by Georgetown students from the Institute for Public Representation.  Here, the district court denies ICE’s motion for summary judgment in my FOIA suit, ordering ICE to do further searches/explain why it can’t and to review its redaction of certain elements of industry guides to detecting counterfeits supplied to ICE.

I particularly appreciated the court’s wry summary of how this dispute began, including “Tushnet explained to ICE that an irreverent parody of a recognized trademark does not infringe because it creates no confusion over the item’s provenance.”  A bit more:

As reported in the Boston Globe, ICE spokesman Daniel Modricker announced that any item that “debas[es] a mascot—and really anything that denigrates a team—is guaranteed to be contraband.” Nestor Ramos, U.S. Agents Tackle Fake Super Bowl Items, Boston Globe, Jan. 31, 2015).… After reading the article, Professor Tushnet immediately wrote to Modricker seeking clarification of ICE’s position on parody merchandise. Modricker doubled down in his reply: “if one logo [disparages] another logo than it would be infringement.”

When pressed further on ICE’s legal basis for seizing parody items, Modricker looped in attorney Joseph Liberta, Chief of the agency’s Criminal Law section. … Liberta attempted to assuage Tushnet’s concerns by noting that ICE, in consultation with agency and Department of Justice attorneys, relies on “potential fair use provisions and federal circuit-specific case law” when determining whether probable cause supports a seizure. He invited Tushnet to submit a FOIA request to obtain more information about the number of counterfeit seizures ICE had made in recent history.  Two weeks later, Tushnet took him up on his offer ….

How do ICE agents determine what to seize? The court explained:

As noted above, the only material that ICE produced in response to Tushnet’s request for training or guidance documents given to ICE agents were 25 instructional guides provided to ICE by various sports leagues and sports apparel companies.  Tushnet finds it is “implausible that ICE has no documents of its own” that instruct officers on how to distinguish counterfeit marks.  The Court does not share Tushnet’s skepticism on this score. It seems entirely logical that ICE would rely on apparel licensers and manufacturers to point out the unique features of their branded clothing, rather than to expend the resources necessary to develop those guidelines internally. 

But that fact makes it quite important that such guides be legally accurate, rather than overclaiming trademark owners’ rights.  It also suggests that Mr. Liberta's initial claims to me about relying on fair use and federal circuit-specific case law were, let's say, unfounded, which is really too bad.

I also partially prevailed on my challenge to the adequacy of ICE’s search, given the disparity in search terms used by field offices, some of which didn’t even use the terms I specifically identified in my request, such as “dilution.”

FOIA Exemption 7(E) authorizes agencies to withhold “records or information compiled for law enforcement purposes [that] would disclose techniques and procedures for law enforcement investigations or prosecutions, or would disclose guidelines for law enforcement investigations or prosecutions if such disclosure could reasonably be expected to risk circumvention of the law.”  This creates a relatively low bar for agencies, who must only show logically how releasing information would create a risk of circumvention.  On the basis of this exemption, ICE redacted over 300 of the 521 pages of the industry guides it released.

The court reasoned: “[o]n its face, ICE’s justification appears sound:  If ICE agents use these guides to distinguish counterfeit goods, revealing the features they look for could help black market manufacturers improve the ‘authenticity’ of their products and potentially avoid detection. This explanation provides a straightforward link between disclosure and potential violations of the law.” Though the court rejected several of my objections, however, it agreed that “there is no legitimate law enforcement purpose in detecting non-counterfeit goods”:

[Tushnet] maintains that some of the material redacted from industry guides might incorrectly characterize clothing as counterfeit when in fact it is a lawful parody. Withholding such material would therefore serve no “legitimate law enforcement purpose” because ICE has no legal authority to seize these items. Tushnet points to a “No Flyers Zone” t-shirt that features the Philadelphia Flyers logo with the Chicago Blackhawks logo imposed over it as one example of an item that was mislabeled as counterfeit in one NHL product guide. The use of the Flyers logo is lawful, according to Tushnet, “because there is no confusion as to whether the Flyers sponsored the shirt.”  In addition, Tushnet presents evidence that ICE has seized other items in this same vein, which suggests a potential misunderstanding within the agency as to what constitutes trademark infringement. The examples offered by Tushnet give the Court pause because 7(E) redactions would be inappropriate if there is no risk that a law could be violated, and successful parodies do not violate trademark laws…. Given the evidence Tushnet has produced and the agency’s apparently exclusive reliance on industry guidance to discern trademark infringement, the Court finds that ICE has not sufficiently justified its 7(E) redactions and that a material factual dispute remains regarding the applicability of this exemption.

Update on my suit against ICE

Today my lawyer Michael Kirkpatrick had oral argument before Judge Cooper of the DDC.  One of the highlights, for me, was the government's lawyer's concession that he couldn't imagine a situation in which "Yankees Suck" would be counterfeit or infringing.  It's good to hear the government say as much, given that this whole thing started with an ICE representative saying that mocking a team was counterfeiting; that at least some of the manuals provided by the leagues to enforcers use parodies as examples of infringing material; and that parodies/critical versions that couldn't possibly cause confusion are, we've found, at least occasionally swept up in seizures.  One thing that more information may help us learn is whether these instances of overreaching against non-counterfeits are accidents or whether they reflect a defect in practices or training.

In which I sue Amazon again

In FTC v. Amazon, the initial opinion was heavily, albeit badly, redacted.  With the able assistance of Paul Alan Levy from Public Citizen, MediaPost and I have moved to unseal the opinion and the documents on which it's based, on the ground that the public's First Amendment interest in access to judicial proceedings outweighs whatever embarrassment the facts might cause Amazon.  Here's our motion to intervene and motion to unseal.

Further results from my FOIA case against ICE

With the assistance of Georgetown's Institute for Public Representation, I sued ICE to get it to answer some questions about why its representative claimed that "X sucks" merchandise would infringe X's trademark rights, at least where X is a sports team.  ICE has made three productions so far, and it's basically a lot of blurry pictures and heavily redacted guides produced by sports teams/apparel companies.  Assuming that the sample I have is at all representative, it seems to me that ICE does seize a lot of plain counterfeits, but it can also seize more questionably.  Is this design on a shirt likely to cause confusion about source?


Blunt Blowin' Bull with bloodshot eyes and marijuana

Certainly, many of the photos showed counterfeits, including counterfeit tags claiming to be genuine NFL etc. merchandise.  With the caveat that the organization of the production makes start/endpoints for different seizures difficult to determine, there was one seizure centering on Seattle Seahawks merchandise that didn't follow this pattern.  There didn't seem to be counterfeit tags--ICE didn't produce photos of tags for this seizure, and their practice is apparently to photograph tags when they might be counterfeits.  Some of this merchandise bore Seahawks logos alone; others involved Seahawks-related images, or a logo plus other positive/fannish commentary.  Many of those shirt didn't look particularly official to me; they might make good examples for discussion of trademark as a pure merchandising right, which ICE is helping to protect.

Superman/Seahawks

Seahawks with list

Match-up with both teams' logos

Go Hawks Re-Pete

I have no idea what's going on here; suggestions very welcome

Why not us?

SanMedica v. Amazon: how many clickthroughs make likely confusion plausible?

In SanMedica v. Amazon, the court initiallyfound enough evidence of confusion from Amazon’s continued use of a trademark in keyword ads (after it had kicked the seller off its platform, but continued to offer competing brands) to deny summary judgment.  However, the court’s initial opinion redacted the percentage of consumers who saw the Amazon ads and clicked through, and the percentage who bought something after clicking through, which meant that it was impossible to understand how the court had applied the governing 1-800 standard, which holds that clickthroughs provide an upper bound on possible confusion.  With the able assistance of Public Citizen, I intervened, and we ultimately agreed to remove a significant amount of the redactions in the opinion and the underlying documents.  I’m pleased to be able to bring you the crucial paragraph in the opinion:

In the present case, there is similar evidence setting an upper limit on how often consumers were lured to Amazon’s website by clicking on the sponsored ads. It is undisputed that during the Advertising Period, approximately 319,000 sponsored ads were generated. Out of those, there were approximately 35,000 clicks on the sponsored ads. The click to impression rate of the sponsored ads is approximately 11 percent. This rate sets the “upper limit on how often consumers really were lured in such a fashion.” Amazon contends that of the “35,253 users that clicked on the ads for SeroVital, only 984 made any purchase at Amazon.com, a measly 3 percent.” Although consumer purchases constitute three percent, the focus is not on the purchase rate but instead on the 11 percent rate that consumers were lured to Amazon’s website. Eleven percent, although a relative small number, is not so insufficient to suggest that there was no likelihood of confusion.

Trademark law takeaway: not great from a traditional perspective—11% as an upper bound is really low, when 15% is a more normal breakpoint.  However, given what’s known about clickthrough rates, most non-Amazon keyword advertisers can probably breathe a little easier.

Update: SanMedica v. Amazon favorably settles

Victory is mine!  Bring me the finest muffins and bagels in all the land!  Public Citizen represented me in a lawsuit to unseal the key information in SanMedica v. Amazon, a case that applied the 1-800 Contacts rule that ad clickthrough rates provide an upper bound to the percentage of consumers who could have been confused online—but sealed the rate it found sufficient to avoid summary judgment.  After the court granted my motion to intervene, the parties agreed to unseal most of the information we sought, including this fact, which is crucial to understanding the decision.  Thanks so much to Scott Michelman and the other lawyers at Public Citizen who brought this about.  More to come when the unsealing occurs and I can at last see the significance of the case for keyword infringement litigation more generally.

In which I win a motion to intervene

Actually, my fine attorneys at Public Citizen have prevailed in my motion to intervene in SanMedica v. Amazon, the case in which the court found sufficient evidence of likely confusion under the 10th Circuit's 1-800 Contacts rule without telling us what that evidence was.  Thank you, Public Citizen!

Hand in glove: ICE and trademark seizures

More tidbits from my FOIA suit against ICE, including the revelation that they did indeed record at least one seizure of disparaging items, “Baltimore sucks” T-shirts.  (Other entries are suggestive, but too unclear to be sure what was seized without details.)  More to follow, but here’s a bit from the ICE Briefing Book on what it called its “NFL News Conference” held in early 2015, which triggered my lawsuit in the first place: 

The NFL news conference on anti-counterfeiting efforts will be co-hosted by Dolores DiBella, NFL’s Counsel. Also as in past years, invitations have been extended to a local CBP [Customs and Border Protection] representative and a local public safety official to have minor speaking roles at the press conference. The primary topic will be the harm to the public, economy and U.S. businesses posed by counterfeit tickets, merchandise and apparel. The event will include only NFL-credentialed media.

Apparently when you buy counterfeit NFL T-shirts, you're stealing American jobs.  How many licensed T-shirts are made in the US, again?

What is ICE seizing?

Early lessons from my FOIA suit against ICE, which has finally resulted in an initial document production: (1) ICE might as well not be keeping records about what it seizes.  The variations include everything from “suspected counterfeit” [particularly informative!], “counterfeit T-shirts,” “counterfeit NFL shirts,” “NFL shirts,” “Buffalo Bills shirts,” “sample,” and other variants I no doubt missed. 

(2) Spelling is very much hit or miss, and it seems likely (given errors in spelling the generic terms accompanying brand names) that the spelling variations are often produced by the people doing the data entry, which means it’s impossible to tell whether the seized suspected counterfeits are near-perfect or the kind of shanzhai misspellings that signal inauthenticity to all but the most careless/least proficient in English.  (Dolfins!  Qowboys!)

(3) They haven’t produced any photos, so it’s impossible to get a sense of the quality of the seized items and, again, whether they are obvious counterfeits.

(4) I’m pretty curious as to why the “Beer pong hat,” and “marijuana hat” were seized, but there’s no explanation in the list I have.

In which I sue the government

So, remember when ICE held a press conference and claimed that disparaging a mark constituted counterfeiting?  I ended up filing a FOIA request to figure out under what circumstances ICE took this position, and since it was denied, I am now--with the wonderful assistance of Georgetown's Institute for Public Representation--suing to have the requested documents released.  We don't actually know very much about the government's anti-counterfeiting operations, and I'm looking forward to finding out more and examining whether the government's working definition of counterfeiting fits with what the statute really says.

(See, now I really do need a "my lawsuits" tag.)

I intervene in SanMedica v. Amazon to disclose clickthroughs

Kind of a busy day.  With the invaluable help of Public Citizen, I filed a motion to intervene and unseal in the SanMedica case, in which the court redacted the number of clickthroughs received by Amazon as a result of its ads using the plaintiffs' trademark, depriving us of the ability to understand the meaning of the case.  Here is the Public Citizen press release:

Secret Data in Trademark Infringement Lawsuit Against Amazon.com Should Be Unsealed, Public Citizen Tells Court Case Decided Based on Sealed Evidence; Key Material Blacked Out in Court Decision May 26, 2015
Contact: Scott Michelman (202) 588-7739 Angela Bradbery (202) 588-7741

SALT LAKE CITY – The First Amendment requires a court to make public the key facts underlying its decision to permit a trademark lawsuit against online retail giant Amazon.com to proceed, Public Citizen argued today in a motion filed in federal district court (PDF) in Utah. The motion, filed on behalf of Georgetown Law professor Rebecca L. Tushnet, an intellectual property expert, seeks to unseal the rationale for the court’s partially blacked-out March 27 decision and see the unedited arguments that led to that decision. In the underlying case, SanMedica International and Western Holdings sued Amazon, alleging that Amazon drew users to its own site by wrongfully continuing to advertise the plaintiffs’ product SeroVital (a dietary supplement) after Amazon stopped offering the product for sale. As a result, according to the lawsuit, Amazon used the product’s ads to acquire customers that should have gone to the plaintiffs. In a decision filed under seal on March 27 and then released to the public on April 15 with key portions blacked out, the district court permitted the case to proceed. The parties settled the same day that the public version of the decision was released.

“The district court’s decision relied heavily on the numbers of ads Amazon ran, the number of customers who clicked on them and the percentage of those customers who bought products from Amazon,” said Scott Michelman, the Public Citizen attorney who represents Tushnet. “But all of the key numbers are blacked out, so neither the public nor future litigants can know what the threshold is for a valid claim. The First Amendment right to access court records exists so that the public can know the law and how the courts arrive at their decisions.” Public Citizen’s motion targets not only the court’s opinion but also a series of court papers that the parties filed with key portions redacted. The parties did not oppose each other’s requests to file under seal, the motion notes, and the district court granted each request quickly and without explanation. The district court’s partially blacked-out ruling described its decision as a “close” one. But its reasoning is obscured by redactions, Public Citizen explains. For instance, the court writes, “[T]he focus is ... on the [redacted] percent rate that consumers were lured to Amazon’s website. [Redacted] percent, although a relative[ly] small number, is not so insufficient to suggest that there was no likelihood of confusion.”

“A key strength of our adversarial legal system is that we can learn the boundaries of the law from past cases,” Tushnet said. “This is an important case for the development of the law. If key facts about the claim can be kept secret, trademark law risks being different for every litigant, which could produce unfair and arbitrary results.” Tushnet also seeks other blacked-out material relevant to the summary judgment motions, including how much money the plaintiffs sought in damages and information about SeroVital itself. Some of the blacked-out material appears to relate to the inner workings of Amazon’s ad-purchasing system; Tushnet is not seeking that material, however, because it was not relevant to the court’s decision. The case is SanMedica International v. Amazon.com in the District of Utah.

Intellectual property lawyer Marian J. Furst of Salt Lake City is co-counsel for Tushnet. The motions seeking to intervene (PDF) and unseal (PDF) the documents are available here.