PLI Media Law conference
RT: There’s been a rapid and somewhat disorienting shift
from a seemingly ever-growing First Amendment freedom of speech to a seeming
indifference to speech and press based claims (as contrasted to religious
freedom claims) in many categories, concentrated in the sudden conservative
abandonment of the commercial speech doctrine. Historically, conservatives
opposed the commercial speech doctrine, which held out the prospect of greater
regulation of commercial speech than of noncommercial speech. However, as
conservatives identified commercial speech, or even just corporate speech and
practices, they’d like to regulate as being too woke, and in line with
populist, anti-monopoly principles, a number of judges, including judges from
varying political backgrounds, have signaled their lack of interest in free
speech claims, whether made against copyright, trademark, false advertising, or
right of publicity lawsuits. Although the Court long ago recognized that
defamation claims need First Amendment safeguards because the state, through
the judiciary, is making rules for speech, it is now unwilling to recognize the
need for similar limits on other causes of action.
I’m going to talk briefly about last term’s Jack Daniels
case—a trademark infringement and dilution case—as well as Elster, argued last
week, in which the Justices appeared inclined to reject a First Amendment
challenge to the refusal to register the claimed mark “TRUMP TOO SMALL” for
t-shirts. Then I’ll talk about the 9th Circuit case Enigma Software
v. Malwarebytes, which allowed a false advertising claim to proceed based on
one software provider’s use of the terms “malicious” and “threat” to describe
its alleged competitor’s software, despite a dissent raising free speech
arguments. I’ll conclude with some general remarks about commercial speech
doctrine after 303 Creative and the pending NetChoice cases.
Trademark: In Jack Daniel’s v. VIP Products, the Court held
that it did not need to apply or even decide the validity of a
speech-protective limit on trademark law designed to prevent trademark owners
from shutting down expressive works that comment on trademarks, known as the
Rogers test. So long as the defendant
was using a similar term as a trademark to brand its own goods (here, dog toys
that parodied Jack Daniel’s well-known whisky bottle), trademark law’s test for
consumer confusion was all that was required. The First Amendment offered no
further protection for this speech, even though the Court did not categorize it
as commercial speech, and it would not have qualified as commercial speech
under ordinary standards. The Court did
not decide whether the First Amendment offered any protection against trademark
infringement claims even for purely expressive noncommercial speech, as every
circuit to have considered the issue had held.
But three Justices wrote separately to express their skepticism that
even purely expressive noncommercial speech involving trademarks merited a
speech-protective test, basing their argument on the fact that the statute did
not specifically outline such a test.
The Court’s conclusion that a defendant’s use as a trademark
obviates the need for any further consideration of the First Amendment makes
sense, if at all, only as a strong form of the rule that commercial speech
receives significantly less protection that other forms of speech, combined
with an incredibly expansive definition of commercial speech. In Jack Daniels,
the Court claimed that the likely confusion test “does enough work to account
for the interest in free expression” in the context of trademark uses. This language, as others have noted, is very
similar to its previous statement in Eldred that fair use and the
idea/expression distinction internal to copyright law provide all the First
Amendment protection necessary in the context of copyright infringement. But in Eldred, the Court followed its
statement by explaining that the Copyright Clause provided a robust foundation
for federal copyright, and that the contemporaneous adoption of the Bill of
Rights and the first federal copyright act showed their compatibility. In Jack
Daniels, the Court said nothing further; its statement remained ipse dixit. Nor
could the court have said more: There is no Trademark Clause and no Founding
Era federal trademark law; indeed, the Court invalidated Congress’s first
attempt, nearly a century later. And
what current courts call “likely confusion” would be unrecognizable to
Founding-era courts—or post-Reconstruction courts—both of which protected
trademark owners in far more limited circumstances: only in cases of “passing
off,” where the defendant’s goods would substitute for purchases of the
plaintiff’s goods. Current confusion doctrine is very different in many ways,
two of the most important being, first, that there is no materiality
requirement—consumers need not care at all about the thing they’re supposedly
confused about, and that itself makes confusion easier to find because people
aren’t careful about things they don’t care about. Second, modern confusion
covers not just confusion over the source of goods or services but, in many
cases, confusion over whether the plaintiff’s permission was required or
whether there is some sort of undefined “affiliation” between the parties. The
Court’s statement that the likely confusion test provides the only necessary
protection for the First Amendment thus lacks either the explanatory apparatus
it offered in Eldred or a basis in text, history, or structure.
Where, then, is the sufficiency of likely confusion from? It
has long been true that the First Amendment assigned no value to “false
statements of fact,” justifying restrictions on deceptive commercial
speech. But, as I’ll discuss more in a
moment, it is possible to turn almost any objection to speech into the claim
that the speech is misleading. Jack Daniels suggests that speech might get no
First Amendment protection against claims labeled “trademark” claims beyond
whatever dubious safety a multifactor likelihood of confusion test may provide—something
that generally requires expensive litigation to work out. The Court also revived Jack Daniel’s dilution
claim, which was not based on any falsity or risk of confusion but on the
theory that being associated with dog poop would “tarnish” the consumer image
of Jack Daniel’s even though consumers understood that there was no relation
between the companies.
Last week’s oral argument in Elster highlighted both the
Justices’ skepticism about free speech claims in the trademark realm and their
lack of interest in the realities of litigation. The Court seems likely to
uphold the prohibition on registering TRUMP TOO SMALL without Trump’s consent,
even though the use isn’t confusing. This will be based on the assertion that
Elster has the right to use TRUMP TOO SMALL on the front of t shirts even
without a registration, and that he couldn’t be prohibited from doing so. But,
if the Court also says there is a legitimate interest in prohibiting
source-indicating uses of an unconsenting person’s name, then that will suggest
a more formalist version of what counts as a trademark use.
Consider for example this argument by Gov’t: “As long as
[Elster] can use the expression and as long as he can obtain the benefits of
trademark registration by choosing a different source identifier to distinguish
his goods from others, he has all he needs [for his free speech]”
This concept of source identifiers is very important: it has
to be somewhat normative if we are preserving his ability to use the
expression. The limit case is when TM owners convince 15% or more of consumers
that any reference to them at all requires permission—the Court seemed
unwilling to deal with that scenario, which fits the facts of a number of the
Rogers cases it distinguished in JDI where the court simply declared that song
titles and movie titles weren’t being used as indicators of source.
Everything here turns on what the Court understands source
identification to mean; if any use that causes any confusion about anything is source-identifying,
then the Court will be allowing huge amounts of speech suppression, but if it
adheres to the idea that trademark, source-indicating use can be objectively
identified and doesn’t exist merely because there’s some kind of confusion,
that can point the way to a workable regime.
My main worry is that the Justices have no interest in the
fact that lower courts have routinely considered the existence of confusion
about something, including about whether the TM owner gave permission for an
accused use, as proof of trademark use.
For example, Justice Jackson said: “trademark is not about
expression. Trademark is not about the First Amendment and … people's ability
to speak. Trademark is about source identifying and preventing consumer
confusion.”
That has to be a normative concept of source identification!
And it’s incoherent without more (if things are confusing, it’s because they express
a message about source or sponsorship; Justice Jackson has to mean that
trademark infringement is not about protected expression—but that doesn’t
really explain what trademark registration is about).
The Court’s indifference to expressive elements of TMs also
conflicts with the Court’s previous claims in Tam, and even Jack Daniel’s, that
TMs themselves often have lots of non-source-indicating expressive meaning:
Jack Daniels isn’t just a piece of information, it is a lifestyle and brand
identity—but then shouldn’t people be able to criticize that identity if they
aren’t making false claims? Court can’t
seem to figure out how to handle that duality between pure information and
brand value.
The Court’s First Amendment jurisprudence in IP may reflect
its current preoccupation with legislative over judicial restrictions on
speech. The legal realists taught us
that an injunction issued by a judge and enforced by the threat of prison is
every bit as much state action as an ordinance passed by a city council. But
the Court today seems much less concerned about courts actually ordering people
not to speak. The Court’s willingness to see threats to speech in the
government’s refusal to register trademarks but not in the government’s banning
the sale of products bearing messages altogether suggests that the imperial
judiciary is alive and well in the First Amendment arena.
Turning to Enigma and false advertising: Commercial speech
doctrine has generally treated false or misleading commercial speech as
completely unprotected by the First Amendment (with some wiggle room for speech
that might be correctable with further disclosures). The attack on commercial
speech doctrine has historically focused on preventing government regulation of
non-false advertising—speech that might encourage unwise consumption of alcohol
or tobacco, for example. Thus, commercial speech doctrine got closer and closer
to strict scrutiny if the government was regulating outside of the false
advertising context, but remained relatively lax if the government was
targeting deception.
But this sharp cliff between types of regulation puts a
premium on being able to distinguish false and misleading speech from speech
that persuades people to do unwise things. This has always been a weak point:
if you defer to the government on what is misleading, then the government can
regulate a lot of commercial speech, even if most people are persuaded rather
than misled. A few Terms back: NIFLA said that these rules didn’t apply to
antiabortion counseling, even with evidence that clients were misled about what
these antiabortion services did. And some people took that as a signal that
mandatory disclosures generally were in trouble, though I was never
particularly persuaded: my own view is that abortion rules are special and the
Court ignores the ordinary First Amendment analysis in abortion cases.
Enigma is a useful test case because it sounds like a
standard false advertising case: the parties allegedly compete (though the
defendant doesn’t agree that they do) and its software labeled the plaintiff’s
software as a potential threat or potential malware. Although the Ninth Circuit
majority acknowledged that there’s a lot of judgment involved in deciding what
is malware, it rejected the defendant’s argument that threat or malware were
nonactionable opinion, in the context of threat defense software. A dissent
argued that this was classic opinion, given the ambiguities in the definition
of malware, which can include stuff that’s just not helpful. And I tend to
think that we should be willing to find falsifiability where the advertiser
claims some kind of expertise in the field, even if we shouldn’t find ordinary
users’ characterizations of something as malware to be potentially defamatory.
But it’s also useful to note that Enigma applied the Lanham Act, which bars
false or misleading commercial advertising generally, rather than legislatively
or administratively identifying a more specific practice. Thus, case by case
adjudication in court is required to find deception. By contrast, a district
court in Colorado last week granted antiabortion activists the right to
advertise “abortion reversal” despite an legislative finding that there is no
such thing—because the drugs at issue could lawfully be administered for other
conditions, the activists had free exercise rights to promote them for abortion
reversal. Of course the situation isn’t exactly the same, but it’s indicative
of where judicial deference is going—to people making religious claims and not
to legislatures making findings of fact.
Compare the question: can one lawfully advertise conversion
therapy to change a young person’s sexual orientation? NJ case a few years back
allowed false advertising claims against the advertiser to proceed. But: If the
government can decide that the promise of conversion therapy is false, can a
different government also decide that it is false and misleading to offer
gender affirming care to trans youth? If you think that the problem is that
this issue is too politicized to allow one side to be labeled false, consider
how many other things have become politicized, and will be politicized if one
party acts. One recent Major area of contestation has been: vegan meat and
dairy substitutes. Meat-producing states have tried to declare that it’s
inherently misleading to use terms like chik’n or sausage, or even oat milk.
Courts so far have struck those down because in the courts’ own views the terms
aren’t misleading—but there are surveys out there that say otherwise for a
nontrivial number of consumers. In addition, the real problem with the vegan
alternatives is that they often have a very different nutritional profile;
there are tragic cases of children who have suffered diseases like kwashiorkor
because their parents were feeding them only vegan alternatives. And the
evidence is very clear that ordinary consumers do not understand the nutritional
differences between almond milk and cow’s milk. Does that mean a ban on the
term “almond milk” is justified?
The usual response is to punt to disclosures: we’ll fix this
problem by telling people the nutritional content is different! Information
overload makes this less effective; the very power of the term “milk” as a cognitive
shortcut tends to lead people to believe they don’t need to know any more,
meaning that they ignore additional information. So in fact you may not be able
to solve the problem with more disclosures. But they’re politically palatable
in ways that bans usually aren’t, so disclosures are a popular solution.
They can also be incredibly burdensome, often not because of
the costs of publicizing the disclosure itself but because of the record
keeping that needs to be done. For example, keeping records to ensure that all
one’s contractors feed cows organic feed can be difficult, as can identifying
the country of origin of each head of cattle if it’s cheapest to just commingle
them in huge lots. And here we arrive at the latest wrinkle in Commercial
speech doctrine generally:
NetChoice: Texas and Florida, explicitly trying to punish
companies perceived as being too liberal, enacted burdensome requirements on
large social media companies. I’m not going to discuss the viewpoint neutrality
mandate, but both states also ordered social media companies to provide
detailed information to users about every moderation decision they made, with
risks of substantial penalties for every error or failure to disclose enough
information to satisfy the states’ AGs. The 11th Circuit struck this
compelled speech down even using intermediate scrutiny, while the 5th
Circuit held it was an unremarkable disclosure mandate regulating commercial
speech. The key move made by the 5th Circuit was to treat all
content policies as commercial speech, which also suggests that the Washington
Post’s editorial policies are commercial speech. Of course, the point was to
make it extremely expensive, burdensome and risky to moderate content, with the
predictable result that companies would do a lot less moderation, so here we
have disclosure that is defended as protecting consumers—so they’ll know what
the actual terms of service are and how they’re enforced—but actually enacted
to change the underlying content policies. Nor are liberals immune from this:
Washington DC’s AG is seeking to get information about individual facebook
accounts to make a claim about its actual policies on covid denial being other
than they were supposedly advertised to be.
I hesitate to make predictions here but I posit to you that
a Court that no longer sees the point of protecting nonreligious speech, or no
longer sees media entities as having their own speech interests distinct from
those of their users, is going to be dangerous for media freedom far beyond
intellectual property claims.
Bennett Cooper (Bad Spaniels/VIP): Without a safe harbor, we
have to fight hard on the confusion arguments on remand. We are confident in
that, as well as a challenge to the viewpoint discriminatory nature of dilution
by tarnishment. You have to sell the joke: if the courts don’t get it, you’re
having a harder time: Kagan didn’t get the joke.
Panel Leader: Jeremy Feigelson: Is the 1A in trouble when
courts avoid doctrine and say “this is just a TM case”?
A: We’ll have to see. Speech as its own category/do other
categories take free speech into account, for example if the multifactor test
can offer true speakers offramps from the multilane highway of litigation, that
can help.
Bruce Johnson: Commercial speech doctrine is becoming more
uncertain as a category. NetChoice cases are compelled editorial transparency;
both 11th and 5th Circuit used commercial
speech/Zauderer, which is the weakest test available. It only applies to ads!
Why use this category to interpret compelled editorial transparency? Baffling!
Unclear to what extent we’ll have commercial speech as a separate doctrine, as
opposed to “Zauderer for speech we don’t like.” AG pressures: efforts to
prohibit Yelp users from communicating about crisis pregnancy centers, using
the Little FTC act in Texas. Washington/Value Village case: if we can squeeze
this into commercial speech, we can regulate the hell out of it. Users were
commenting about things protected by 1A until Ken Paxton decided that wasn’t
ok. Making commercial speech law incoherent.
Jennifer Rothman: Commercial speech doctrine has been
eroded/diluted over decades. Part of a larger story of Court deferring to
property rights over speech. After Jack Daniels we have questions and concerns
both within TM and 1A. Rogers is important to a variety of creative industries
that have relied on it for references to identities/marks. Court clearly
signals that it doesn’t love Rogers, but they didn’t make a bright line about
use as a mark. But that’s also true of Rogers itself, where Ginger Rogers’ name
was incorporated into the movie title itself, which seems to function at least
in part like a source identifier (contingent on identifying a particular
creative work, which because of copyright law will be from a particular source
for a while).
Thinks the lower courts will still gravitate to Rogers in
traditional expressive use cases, not product cases. But products can be
expressive, some will say! Still, courts are going to draw lines. Not really a
comment; using value of mark to make a joke, which is generally disfavored. We
may see less destabilization b/c courts are still going to prefer/protect
traditional expressive uses and not “commercial products.” But this also
destabilizes defenses to ROP claims, in conjunction with Warhol, b/c some courts
have also relied on Rogers in ROP cases. To the extent that Warhol calls
transformativeness into question v. market harm, states like California that
use transformativeness in 1A defenses to ROP will also see those destabilized.
Finally, dilution: doesn’t think Court shares her (or RT’s) concerns—dicta
suggests that Court will find dilution by tarnishment ok, even if it seems
viewpoint-based.
Johnson: Bigelow v. VA was an abortion speech case; Court
went out of its way to protect commercial speech where something was legal in
one state and illegal in another; we’ll see more efforts pushed on whether and
to what extent speech is purely commercial or whether there are expressive
rights underneath it.
Feigelson: connection to 303 Creative?
A: decided on slender record; Court went out of its way to
protect website developer from being compelled against her will to respond to a
person who allegedly called her and wanted to have a gay wedding website; this
doesn’t seem to be true but the Court still said she didn’t have to provide
service. Court eliminated Colorado’s disclosure requirements too. Zauderer
requires you to provide corrective speech where your affirmative commercial
speech would otherwise be misleading; but the test is “purely factual/noncontroversial,”
and that doesn’t seem to apply here.
Feigelson: Elster: is Court going out of its way to limit
the First Amendment?
A: not really what the case is about—don’t overcommit to
preliminary interpretations. But they were suggesting that the free speech
interests weighed in the other direction: allowing one registration of Trump
Too Small could prevent other criticism of Trump b/c Elster could block
competitors. There are some arguments that the 1A weighs against registration
here [though will the Court recognize that with other registrations?]. Other
connections: 2(c) bar is about gov’t interest in protecting individual identity—“longstanding”
protections for individual identities in TM law. This is an instance where you
don’t have to show likely confusion about sponsorship/endorsement. That
suggests again the Justices’ willingness to allow claims not rooted in showings
of likely confusion—which would be helpful for dilution’s survival.
Cooper: These cases are on a spectrum: viewpoint
discrimination v. neutrality; whether you’re talking about the right to have a
good name v. effectiveness as a source identifier. Tam/Brunetti found viewpoint
discrimination, whereas 2(c) is viewpoint-neutral [except in combination with
2(a) and 2(d)]. Tarnishment situation is more like Tam/Brunetti, because you’re
talking about banning nonconfusing uses with a particular viewpoint. 2(c)’s
consent requirement is more about source identification: worry about potential
endorsement/connection. Blurring is about effectiveness of mark as wayfinder in
marketplace; the Court might favor that b/c it’s closer to TM’s function.
Tarnishment is a “happy talk” provision.
Turning to ROP:
Rothman: Writers’ strike is settled; actors continue—they
seem close on residuals, pay structure, streaming. But still outstanding in
external reporting is AI provisions. Concerned about studios/producers scanning
actors and reusing them w/out control or payments.
Lots of legislation proposed furthering particular parties’
interests, including NO FAKES act—resurrecting James Dean—deceased performers
could replace work for the living, and we don’t necessarily want to create more
robust rights in the dead to allow that replacement to take place. SAG/draft
legislation hasn’t addressed need for disclaimers/revealing to the public that
the actual person had nothing to do with the work, which could be important far
beyond creative industries.
Originally NO FAKES was fast-tracked, sponsored across party
lines. But it turned out not everyone could agree on it.
Feigelson: should Enigma cause worries about product
reviewers?
RT: depends on how integrated you are with your sponsors. If
you are heavily integrated, then disclosure at a minimum is required. And if
you’re making health claims then yes you should worry. But Wirecutter is
unlikely to be treated as commercial speech—that’s a product review.
Johnson: “threat” and “malware” would be 1A protected
opinion in most circumstances; the 9th Circuit found them suspicious
b/c a competitor used those terms—an exception created when it’s not quite
commercial speech (footnote says that it wasn’t clear this was commercial
speech; that was for remand). Liability wouldn’t arise in a pure op-ed.
