STATE OF WASHINGTON v. META PLATFORMS, INC., formerly doing business as FACEBOOK, INC.
No. 84661-2-I
IN THE COURT OF APPEALS OF THE STATE OF WASHINGTON DIVISION ONE
FILED 12/2/2024
DÍAZ, J.
PUBLISHED OPINION
The State sued Meta Platforms, Inc. (Meta) for violating relatively new provisions of Washington‘s Fair Campaign Practices Act and its implementing legislation, namely,
The superior court granted the State‘s motion for summary judgment on both liability and damages, denied Meta‘s cross-motion, and entered a $35 million judgment against Meta, two-thirds of which consisted of a civil penalty and one-third of which was an award of the State‘s attorney fees and costs.
Meta appeals and argues that the disclosure law violates the First Amendment to the United States Constitution and is preempted by the federal Communications Decency Act,
I. BACKGROUND
A. Overview of the Factual Background
The parties do not dispute the following factual background. Meta is the parent company of Facebook, Inc., and other international online social media networking platforms, serving over 2 billion monthly users. In pertinent part, Meta allows people and organizations to purchase advertisements directed to Meta‘s general membership or selected parts of their membership. These advertisements have included political topics, such as promoting candidates for local and state elections. As will be discussed in more detail below, Meta allows advertisers to target their ads to specific users or groups of users, encompassing a wide range of demographic traits, through a self-service tool. These demographic traits include Meta users’ age, race, ethnicity, gender, location, interests, and more.
In approximately May 2018, Meta created an “Ad Library.” The Ad Library includes advertisements Meta identifies as “political,” which it commits to retain for
In June 2018, the State sued Meta for the first time for failing to comply with the disclosure law. In its complaint, the State alleged that at least two members of the public requested information from Meta about political ads hosted on its platforms, and Meta did not provide the information the disclosure law required.2 In December 2018, Meta entered into a stipulated judgment, where it agreed to pay $200,000 in damages.
On December 28, 2018, Meta announced that it would no longer accept ads related to Washington electoral campaigns. Despite the ban, however, ads related to political campaigns in Washington still appeared on Meta‘s platforms and in its Ad Library. For example, through Meta‘s self-service portal, advertisers placed approximately 1,600 ads related to Washington‘s 2019 elections on Meta‘s platforms.
In 2019, two requesters observed political ads on Meta‘s platforms and contacted Meta by email, requesting information about the ads pursuant to the disclosure law. Both requestors filed complaints with the Public Disclosure
Later and separately, a third requestor also viewed political ads on Meta‘s platforms, but did not see the same ads in Meta‘s Ad Library. Between 2019 and 2021, this requestor unsuccessfully attempted to request information about various ads from Meta. Unlike the previous requestors, this person also visited Meta‘s facilities in-person, requesting to review their records. Unsatisfied, the requestor filed a complaint with the AGO, who forwarded it to the PDC, which returned the matter to the State to litigate.
B. Superior Court History
In April 2020, the State filed a complaint in the King County Superior Court. The State alleged, among other things, that Meta hosted ads on its platforms, but did not provide full or timely information to the three requestors in violation of the disclosure law.
After substantial discovery, including depositions of the parties’ experts, Meta and the State each moved for summary judgment in July 2022. As told by the trial court, “Meta has now [asked] the Court to find the statute and regulations unconstitutional, and the State has brought a cross motion for enforcement of those elements of the law.”
The superior court granted the State‘s and denied Meta‘s motion for summary judgment in October 2022. The superior court also granted the State‘s motion to enter judgment against Meta, awarding $24,660,000.00 in civil penalties
C. Appellate History
Meta appeals from the orders on both the summary judgment motions and the judgment, as well as the injunction embedded in each. In November 2022, Meta filed an emergency motion for a stay of the superior court‘s injunction before a commissioner of this court. The commissioner granted Meta‘s motion for a stay of the injunction, concluding that reasonable minds could differ about the application of new relevant federal caselaw, and denied the State‘s motion for a bond.3 4
II. ANALYSIS
As this is an appeal of the orders on summary judgment, we review whether “the pleadings, depositions, answers to interrogatories, and admissions on file, together with the affidavits, if any, show that there is no genuine issue as to any material fact and that the moving party is entitled to a judgment as a matter of law.”
“Summary judgment ‘is subject to a burden-shifting scheme.‘” Welch v. Brand Insulations, Inc., 27 Wn. App. 2d 110, 114, 531 P.3d 265 (2023) (quoting Bucci v. Nw. Tr. Servs., Inc., 197 Wn. App. 318, 326, 387 P.3d 1139 (2016)). “The moving party meets its initial burden by submitting evidence demonstrating that it is entitled to a judgment as a matter of law.” Bucci, 197 Wn. App. at 326. “The burden then shifts to the nonmoving party to set forth ‘specific facts which sufficiently rebut the moving party‘s contentions and disclose the existence of a genuine issue as to a material fact.‘” Id. (quoting Ranger Ins. Co. v. Pierce County, 164 Wn.2d 545, 552, 192 P.3d 886 (2008)) (emphasis added).
Summary judgment gauges whether the nonmoving party has met their
Meta challenges the disclosure law only as it violates its First Amendment rights and because it is preempted by Section 230. We address each in order.
B. First Amendment
“[T]he Government may not suppress political speech on the basis of the speaker‘s corporate identity.” Citizens United v. Fed. Election Comm‘n, 558 U.S. 310, 365, 130 S. Ct. 876, 898, 175 L. Ed. 2d 753 (2010).
1. Overview of the Disclosure Law
Given the unique nature of Washington‘s disclosure law, and the gravity of the various constitutional interests at stake in this case, we first endeavor to provide the necessary historical and procedural framework of the challenged laws.
a. Origins of the Disclosure Law
More than 50 years ago, a bipartisan coalition of nonprofit organizations, political officials, and prominent media figures formed a group which drafted Initiative 276 (I-276). HUGH A. BONE & CINDY M. FEY, THE PEOPLE‘S RIGHT TO KNOW: AN ANALYSIS OF THE WASHINGTON STATE PUBLIC DISCLOSURE LAW 1-2 (1978). The
The promotional materials for I-276 set forth its purpose of promoting the people‘s “right to know” by providing the public with a greater access to information about elections, stating: “Our whole concept of democracy is based on an informed and involved citizenry. Trust and confidence in governmental institutions is at an all time low. . . . Initiative 276 brings all of this out into the open for citizens and voters to judge for themselves.” State of Washington Voters Pamphlet, General Election 10 (Nov. 7, 1972). As later summarized by our Supreme Court, I-276 sought “to enlarge the information based upon which the electorate makes its decisions. The right of the electorate to know most certainly is no less fundamental than the right of privacy.” Fritz v. Gorton, 83 Wn.2d 275, 298, 517 P.2d 911 (1974) (plurality opinion).
In 1972, voters approved I-276 with a 72.01% margin, carrying every county in Washington. BONE & FEY, supra, at 3. I-276 “established the PDC and formed
In 1973, Washington enacted the Fair Campaign Practices Act (FCPA), which codified I-276. Evergreen Freedom Found., 192 Wn.2d at 790; LAWS OF 1973, ch. 1. In adopting the purposes of I-276, the FCPA declares as the public policy of Washington:
(1) That political campaign and lobbying contributions and expenditures be fully disclosed to the public and that secrecy is to be avoided. . . .
(5) That public confidence in government at all levels is essential and must be promoted by all possible means. . . .
(10) That the public‘s right to know of the financing of political campaigns and lobbying and the financial affairs of elected officials and candidates far outweighs any right that these matters remain secret and private.
LAWS OF 1973, Ch. 1, § 1;
In enacting the disclosure law, the legislature gave a twice repeated command that its provisions shall be “liberally construed” to effectuate its policies—“to promote complete disclosure of all information respecting the financing of political campaigns and lobbying . . . so as to assure continuing public confidence in fairness of elections and governmental processes, and so as to
b. Substantive Requirements of the Disclosure Law
With the above historical and statutory context in mind, we turn to what the disclosure law currently requires in pertinent part.6
A commercial advertiser is “any person that sells the service of communicating messages or producing material for broadcast or distribution to the general public or segments of the general public . . . used for the purpose of appealing, directly or indirectly, for votes or for financial or other support in any election campaign.”
The statute directs that “The documents and books of account” in question must include: “(a) The names and addresses of persons from whom it accepted political advertising or electioneering communications; (b) The exact nature and extent of the services rendered; and (c) The total cost and the manner of payment for the services.”
The legislature further delegated authority to the PDC to implement the statute by promulgating and defining through rule making what it means to “maintain current books” and “make open for inspection.”
In 2018, the PDC updated its existing regulations for disclosures by “digital communications platforms,” such as Meta,7 which are a subset of “commercial advertisers.”
(a) The name of the candidate or ballot measure supported or opposed;
(b) A copy of the advertisement or communication;
(c) The name and address of the sponsoring person or persons actually paying for the advertisement or electioneering communication;
(d) The total cost of the advertising or electioneering communication; and
(e) The date(s) the commercial advertiser rendered service.
Moreover, the rule requires that digital communications platforms in particular must maintain in their books and be prepared to disclose “[a] description of the demographic information, the statistical characteristics of a population (e.g., age, gender, race, location, etc.), of the audiences targeted and reached, to the extent such information is collected by the commercial advertiser as part of its regular course of business, and the total number of impressions generated by the advertisement or communication.”
Meta argues that the foregoing provisions violate its First Amendment rights and begins by claiming that we must examine the disclosure law with strict scrutiny.9
2. Applicable Level of Scrutiny
Regulations that burden political speech must typically withstand strict
That said, “unlike,” for example, “limits on election-related spending10—election-related disclosure and disclaimer requirements ‘impose no ceiling on campaign-related activities.‘” Gaspee, 13 F.4th at 85 (quoting Citizens United, 558 U.S. at 366) (emphasis added). In other words, “disclosure requirements are constitutional because they ‘d[o] not prevent anyone from speaking.‘” McConnell v. Fed. Election Comm‘n, 540 U.S. 93, 201, 124 S. Ct. 619, 693, 157 L. Ed. 2d 491 (2003), overruled on other grounds by Citizens United, 558 U.S. at 310 (quoting Brief for FEC in Opposition in No. 01-582 et al. (DC), p. 112). This is so because, “disclosure and disclaimer regimes are cut from different cloth” than other burdens on political speech. Gaspee, 13 F.4th at 85. Both federal and our state courts have explored what this distinction means for purposes of the intensity of the review we apply.
Federal courts have applied exacting scrutiny to election-related disclosure requirements for decades. Buckley v. Am. Constitutional Law Found., Inc., 525 U.S. 182, 199, 119 S. Ct. 636, 645, 142 L. Ed. 2d 599 (1999) (considering
Washington courts also have applied exacting scrutiny to Washington registration and disclosure requirements, including for political committees. State v. Grocery Mfrs. Ass‘n, 195 Wn.2d 442, 461, 461 P.3d 334 (2020) (Grocery Mfrs. Assn. I) (regarding a trade association‘s registration and disclosure requirements); Evergreen Freedom Found., 192 Wn.2d at 799.
Most recently, in Evergreen Freedom Found., our Supreme Court examined separate provisions of the FCPA, which required the disclosure of financial information of entities providing pro bono legal services in furtherance of efforts to influence elections. There, the Evergreen Freedom Foundation “created sample municipal ordinances and ballot propositions for citizens to use to advance certain
Here, the parties do not dispute that the law at the center of this appeal is a disclosure requirement. Nonetheless, Meta argues that we should examine the disclosure law under strict scrutiny for two reasons, neither of which is persuasive.
First, relying on Reed v. Town of Gilbert, 576 U.S. 155, 163, 135 S. Ct. 2218, 2226, 192 L. Ed. 2d 236 (2015), Meta avers that the disclosure law warrants strict scrutiny because the statute “is a content-based regulation of speech since it ‘applies to particular speech because of the topic discussed‘—namely, Washington politics.” (Emphasis added).
This argument flies in the face of each federal and state case reviewed herein, all of which dealt with “political topics” and none of which deemed such a disclosure to be “content-based” regulation. This conclusion is particularly apt here because, unlike in Reed, where the challenged statute “single[d] out specific subject matter for deferential treatment,” the disclosure law attaches potential liability to any advertisement categorized as “political” without regard to an ad‘s specific partisan or other content. Reed, 576 U.S. at 163;
Meta, however, offers no authority in support of the proposition that candidates and non-candidates’ disclosure obligations should be analyzed differently. On the contrary, the Supreme Court in Bonta held that “[r]egardless of the type of association, compelled disclosure requirements are reviewed under exacting scrutiny.” 594 U.S. at 608. What‘s more, our Supreme Court applied exacting scrutiny to the claims presented by Evergreen Freedom Foundation, which was a firm that provided pro bono legal services and was not an electoral candidate. Evergreen Freedom Found., 192 Wn.2d at 801.
Further, third-party platforms are indeed strongly motivated to comply with disclosure laws; namely, in order to continue receiving advertising revenue from candidates and political action committees and to avoid sanction. The television stations who broadcast political advertisements, discussed in Citizens United, 558 U.S. at 366-67, provide proof of such motivation, despite the disclosure requirements imposed upon them.12
3. Exacting Scrutiny Analysis
“[A] campaign finance disclosure requirement is constitutional if it survives exacting scrutiny,” which means that the disclosure requirement is (a) “substantially related to a sufficiently important governmental interest[,]” Brumsickle, 624 F.3d at 1005, and (b) “narrowly tailored to the interest it promotes.” Bonta, 594 U.S. 611 (citing Shelton v. Tucker, 364 U.S. 479, 81 S. Ct. 2471, 5 L. Ed. 2d 231 (1960)). In other words, as Meta agrees,13 exacting scrutiny is a two-step analysis: first, whether there is a sufficiently important government interest that bears a substantial relationship to the statute and then whether the disclosure law is sufficiently narrowly tailored. Gaspee, 13 F.4th at 96.
a. Substantially Related to a Sufficiently Important Government Interest
We hold, as a matter of law, that
i. Sufficiently Important Government Interest
The State asserts various governmental interests advanced by the required disclosures in the challenged law. We focus only on one: the need to timely inform the electorate about who is expending money to influence an election in our state and how that money is being spent.
Both federal and state courts have long recognized the importance of fostering an educated electorate. The United States Supreme Court in Citizens United held that laws requiring disclosure of campaign expenditures enable “the electorate to make informed decisions and give proper weight to different speakers and messages.” 558 U.S. at 371. Further, in McConnell, the Court ultimately upheld one of the requirements for timely disclosure of advertiser data considered there because, “[g]iven the relatively short timeframes in which electioneering communications are made, the interest in assuring that disclosures are made promptly and in time to provide relevant information to voters is unquestionably significant.” 540 U.S. at 200. The Court held that disclosure requirements “perform an important function in informing the public about various candidates’ supporters before election day.” Id. at 201. Even more broadly, the Court upheld another disclosure requirement because “the important state interests that prompted the Buckley Court to uphold [the Federal Election Campaign Act]‘s disclosure requirements—[including] providing the electorate with information—apply in full to [the Bipartisan Campaign Reform Act].” Id. at 196.
(1) That political campaign and lobbying contributions and expenditures be fully disclosed to the public and that secrecy is to be avoided. . . .
(10) That the public‘s right to know of the financing of political campaigns and lobbying and the financial affairs of elected officials and candidates far outweighs any right that these matters remain secret and private. . . .
Evergreen Freedom Found., 192 Wn.2d at 790-791 (some emphasis added) (quoting LAWS OF 1973, ch. 1, § 1 and
With that intent in mind, and with respect to the kinds of disclosures at issue here, “Political Advertising and Electioneering Communications,” our legislature explicitly found that “the state has a compelling interest in providing voters information about electioneering communications in political campaigns concerning candidates for state, local, or judicial office so that voters can be fully informed as to the: (i) Source of support or opposition to those candidates; and (ii) identity of persons attempting to influence the outcome of state, local, and judicial candidate elections.”
In turn, Washington courts have consistently held that the State has an important interest in ensuring transparency in elections as “the public, acting as legislators on ballot propositions . . . has the right to know who is lobbying for their votes . . . [a]s such, courts have repeatedly recognized a sufficiently important government interest in as-applied challenges to campaign finance disclosure laws.” Grocery Mfrs. Assn. I, 195 Wn.2d at 463 (citing Brumsickle, 624 F.3d at 1006-07).
When courts have considered and upheld election-related disclosure laws under the exacting scrutiny standard, they stress—often in sweeping terms—that transparent elections not only further an important government interest but sustain democracy itself. See, e.g., Buckley v. Valeo, 424 U.S. 1, 14-15, 96 S. Ct. 612, 632, 46 L. Ed. 2d 659 (1976) (“In a republic where the people are sovereign, the ability of the citizenry to make informed choices among candidates for office is essential“); Reed, 561 U.S. at 199 (“[p]ublic disclosure also promotes transparency and accountability in the electoral process to an extent other measures cannot“); Gaspee, 13 F.4th at 95 (“a well-informed electorate is as vital to the survival of a democracy as air is to the survival of human life.“).
The State meets its initial burden of establishing a “sufficiently important governmental interest,” and Meta does not seriously challenge that an educated
Thus, we hold the State has established that an educated populace is a “sufficiently important governmental interest,” which Meta has not offered any evidence to rebut. Id.
ii. Substantial Relationship
We next consider whether the requirements of the disclosure law bear a “substantial relationship” to that important governmental interest.
In support of its motion for summary judgment, the State offered testimony that it was necessary for Meta to provide information such as the advertiser‘s address, target audience, and payment method to capture the fullest picture of an advertiser‘s identity and intent. It so met its initial burden on summary judgment.
In response, Meta argues there is no substantial relationship between the government‘s interest in an educated electorate and the disclosure law because some provisions of the statute require disclosure of unnecessary information, such as the payment method of the advertisement. We disagree.
Meta offers no authority that this type of analysis—where it attempts to cherry pick certain provisions—is the proper way to analyze a statutory scheme or can be used to undermine the overall substantial connection between the disclosure law and its purpose. State v. Loos, 14 Wn. App. 2d 748, 758, 473 P.3d 1229 (2020) (“When a party provides no citation to support an argument, this court will assume that counsel, after diligent search, has found none.“). On the contrary, when interpreting regulations, we construe the act as a whole, giving effect to all of the language used. Durant v. State Farm Mut. Auto. Ins. Co., 191 Wn.2d 1, 8, 419 P.3d 400 (2018).
Meta next argues the disclosure law‘s requirements do not further the government‘s goal of educating the electorate because, according to Meta, only three people requested information from Meta under the disclosure law. Meta also asserts the information the disclosure law requires is already available by other means, whether its own Ad Library or disclosures required by other laws. These arguments fail for two important reasons.
First, Meta cites no authority that a certain number of voters or only individual voters must use the disclosure law to inform their own personal vote. Indeed, we are aware of no other covered entity that has made an argument on de minimis use by voters. Further, as the voting rights amici argue, it is commonly
Similarly, and second, Meta offers no authority for the proposition that a given law may not be partially duplicative of any other law, or it risks lacking a substantial relationship to the established interest. And there is no evidence in the record that all the information required by the disclosure law is available by other means. While it is true that some voters may obtain some information about who purchases advertisements via a given candidate‘s disclosures or even Meta‘s Ad Library, there is no dispute that Meta and only Meta retains all the information required by the disclosure law including, e.g., the method of payment.
By way of further context, practices such as the micro-targeting of advertisements means no one online audience member can know the full range of advertisements by a candidate or organization because the algorithm may filter ads away from or to a voter based on their personal preferences or demographics. In other words, only Meta holds the information that would answer the public‘s questions about who is purchasing ads on behalf of candidates and organizations, how, and for what purpose. The voting rights amici emphasize this point, asserting:
These unique features of digital advertising pose new threats to democracy. The practice of micro-targeting means that online
audiences have little understanding of the full range of advertising run by a candidate or advocacy group, including the different messages other voters are being shown. This new ability to secretively direct a range of specially tailored, and perhaps even conflicting, messages to different audiences is incompatible with the core legitimizing aspects of democratic society—such as “publicity and transparency for the deliberative process.”
(Quoting Jürgen Habermas, Political Communication in Media Society: Does Democracy Still Enjoy an Epistemic Dimension? The Impact of Normative Theory on Empirical Research, 16 COMM‘NS STUD. 411, 413 (2006)).
Remarkably, as explained by State‘s expert, Dr. Shannon McGregor, a professor of journalism and media at the University of North Carolina, even advertising agencies and campaigns themselves do not have access to the full targeting data, meaning it would not be entirely captured in the campaign‘s own disclosures that are required by other laws. In full, Dr. McGregor explained:
[S]ay I want to target people with an interest in native gardening with this ad. And [Meta] sends it out . . . to some people. But if it seems to be doing really well with white women in particular areas, there are other type of native gardener interests. But it seems to be doing particularly well there.
Then the [Meta] ad auction system, the way I understand it when it‘s being explained to me by [Meta] employees who work on it, is that when those people who engaged or expressed interest in it will get more of it. Right? Because, like, more people like them – right – will get more of it. And that will help ensure that the ad is both seen by lots of people, but by people especially who want to engage with it. That‘s the benefit. Right? For the – for [Meta], in terms of engagement. And it‘s a benefit for the ad buyer in terms of they have engaged people, right, in their ads.
But that changes who‘s being targeted. Right? Because there‘s like the initial targeting list. And then it‘s like how that‘s shaped by the auction. And then that‘s when we end up with the final reach or impression numbers. And I think that middle part is not really known,
to my knowledge, to the consultancy or campaign placing the ad. What change that made.
Dr. McGregor summarized by explaining, “if a campaign or consultancy uploads a lookalike file for a lookalike audience to be created, based on, they have that information.” However, “they will never have the information about the actual targets of that ad campaign that was based on the lookalike audience that was created by [Meta].”
The evidence in the record, including Dr. McGregor‘s sworn testimony, shows that the disclosure law‘s “requirements specify information about individual political ads that is readily available to advertising platforms, but may not be available to campaigns.” Indeed, Meta itself stresses this advantage to potential advertisers because Meta collects such detailed information about its members. Clerk‘s Papers (CP) at 7163-65 (“Your audience is on Facebook“). Meta otherwise does not put forth evidence that there is not substantial relationship between, on the one hand, the disclosure law and, on the other hand, an informed electorate. The only evidence in the record supports the State‘s position on this point.
Thus, Meta has not created a genuine issue of material fact that the disclosure law does not bear a substantial relationship with the government‘s interest in fostering an informed electorate.
b. Sufficiently Narrowly Tailored
We hold that Meta did not create a genuine issue of material fact that
i. No Genuine Issue of Material Fact as to Unnecessary Burdens on Meta‘s First Amendment Rights
“While exacting scrutiny does not require that disclosure regimes be the least restrictive means of achieving their ends, it does require that they be narrowly tailored to the government‘s asserted interest.” Bonta, 594 U.S. at 608 (emphasis added). “[A] reasonable assessment of the burdens imposed by disclosure should begin with an understanding of the extent to which the burdens are unnecessary, and that requires narrow tailoring.” Id. at 611 (emphasis added). In other words, the “[t]ouchstone for exacting scrutiny [under the First Amendment] is whether there is a ‘fit [between the chosen means and the purported ends] that is not necessarily perfect, but reasonable.‘” Washington Post v. McManus, 944 F.3d 506, 523 (4th Cir. 2019) (emphasis added) (quoting McCutcheon v. Fed. Election Comm‘n, 572 U.S. 185, 218, 134 S. Ct. 1434, 188 L. Ed. 2d 468 (2014)).
In Bonta, for example, the Supreme Court concluded that a mandatory state collection of charitable donation forms was not sufficiently narrowly tailored to offset the potential privacy concerns of donors. 594 U.S. at 615-16. The Court concluded the mandatory collection requirement compromised donors’ First Amendment right of free association. Id. at 616-17. Meta makes no such argument here. More relevantly, the Court was unconvinced the government needed to collect that quantity of data to combat charity fraud. Id. (characterizing the requirement as “indiscriminately sweeping up the information of every major
Thus far, Washington cases applying exacting scrutiny to other disclosure laws have not examined, in detail, what constitutes a “narrow tailoring” and “unnecessary burdens.” Bonta, 594 U.S. at 611.
Again, “[t]he party moving for summary judgment bears the initial burden of showing that there is no disputed issue of material fact.” Haley v. Amazon.com Servs., LLC, 25 Wn. App. 2d 207, 216, 522 P.3d 80 (2022). “The burden then shifts to the nonmoving party to present evidence that an issue of material fact remains.” Id. We will conduct this step of the analysis with reference to the superior court‘s granting of the State‘s motion for summary judgment.16
(a) The State‘s Initial Showing
At its most general level, the State argues in briefing that the disclosure law is sufficiently narrowly tailored to achieve its purpose of informing the electorate because the “transparency” it generates “assists the public in understanding the context for why a group is advocating for or against a position in an election.” In other words, “[t]he law provides the public ‘with the information with which to assess the various messages vying for their attention in the marketplace of ideas.‘” Br. of Resp‘t at 47 (quoting Brumsickle, 624 F.3d at 1008).
To explain why it is necessary to disclose information identifying all the discrete pieces of data the disclosure law requires, Dr. McGregor explained:
The value ad that Meta has is the ability to target along very specific
lines that is different than other types of advertising. In different mediums.
And when asked what the “basis” for making that assertion was, Dr. McGregor responded:
Because that is the reason that people advertise on Meta and other digital platforms. Is because it is lower cost and more precise.
According to the State, “[t]argeting is the primary service of a product rendered by Meta for sponsors of political ads and is essential to understanding the purpose and meaning of political ads appearing on that platform.” And each piece of data required by the disclosure law supports the requestor‘s understanding of the nature of the targeting involved in a given ad.
Thus, we hold that the State has put forth competent evidence that there is, not just a reasonable “fit” with the government‘s goals, but that the goal is perfectly aligned with the data required by
Moreover, the State argues that determining whether the content of an ad is “political” is not an unnecessary burden, or as challenging as Meta claims, because “Meta already knows how to and does identify political ads based on a definition.”17 Specifically, the State avers Meta is not burdened unnecessarily by the disclosure law, but is simply refusing to filter the ads that it already tracks
In support, the State proffers documents related to the online “Meta Business Help Center,” which previously defined “[a]ds about social issues, elections or politics” as:
- Made by, on behalf of, or about a candidate for public office, a political figure, a political party, a political action committee, or advocates for the outcome of an election to public office; or
- About any election, referendum, or ballot initiative, including “go out and vote” or election campaigns; or
- About social issues in any place where the ad is being placed; or
- Regulated as political advertising.
Finally, as another State expert, Dr. Laura Edelson, a postdoctoral researcher an New York University whose expertise is in Facebook political advertising, testified, Meta‘s alternate path to providing the same data (training a new sorting model, which will be discussed further below) was “[d]efinitely one of the more expensive options.” Dr. Edelson recommended, instead, that Meta “set up Washington state as its own region within the Ad Library,” which again is already collecting ad data and would cost approximately $200,000. In arriving at this estimate, Dr. Edelson reasoned that “to make the change itself would probably – you need one engineer – I‘m going to be really, really conservative and say a quarter. And obviously, you need to – you know, a few other people like a tester; you need a product manager to sort of oversee things.” Regardless of the precise cost, Dr. Edelson opined that this option would be a “very inexpensive route to comply with the Washington State [disclosure law].”
Dr. Edelson‘s recommendation was based in part on her own experience “buil[ding] a topic model for elections that covered state and local elections,” but not specifically for Washington, and on the undisputed fact that Meta has already altered its Ad Library to accommodate the election laws of other nations, for example, the United Kingdom and Canada. Dr. Edelson explained that “[Meta] made additional data about the advertiser, including the address, available through the ad library, and I believe to be in compliance with -- with U.K. law” around the time of the 2019 election.
Dr. Edelson testified in detail as to Meta‘s assertion of an unnecessary burden and concluded that Meta “can comply with these regulations” as it is “[s]omething that is well within—very comfortably within Meta‘s capabilities; it wouldn‘t strain it as a business.” We hold the State has met its initial burden to show a “fit” that is neither ill tailored to the government‘s goal or otherwise unnecessarily burdensome on Meta‘s rights.
(b) Meta‘s Response
As an initial response, Meta argues the State‘s assertion that it could easily comply, such as by modifying the existing Ad Library, were “premised on guesswork and erroneous assumptions.” Such a “conclusory” statement does not defeat the foregoing evidence offered by the State at this stage of the summary judgment burden shifting. Grimwood v. Univ. of Puget Sound, Inc., 110 Wn.2d 355, 359-60, 753 P.2d 517 (1988), abrogated on other grounds by Mikkelsen v. Pub. Util. Dist. No. 1 of Kittitas County, 189 Wn.2d 516, 527, 404 P.3d 464 (2017).
More generally, Meta argues that the “marginal benefit” of the disclosure
First, Meta asserts it is indeed burdensome to determine which ads are political. As told by Meta, the disclosure law makes identifying ads challenging because the term “political advertising” broadly includes any ad “used for the purpose of appealing, directly or indirectly, for votes or financial or other support or opposition to any election campaign.” Meta avers the challenge is compounded by the term “electioneering communication,” which it asserts covers any ad that “clearly identifies a candidate” even if it does not include the candidate‘s name. According to Meta, determining whether an ad meets the above definition “often requires not just [1] initial screening by an algorithm, but also [2] subsequent human review.”
As to the human review argument, Meta cited to the testimony of its employee, Frank Antzoulis, a program manager, describing how Meta‘s “markets team” conducted “sweeps” to “identify any ads that our normal models couldn‘t catch.” In an attempt to comply with the disclosure, Mr. Antzoulis testified that “those sweeps were performed at a biweekly cadence,” although the sweeps later were reduced from biweekly to monthly.
Mr. Antzoulis explained that, as part of their research, the teams identify which candidates are running in Washington, then check if they appear in ads by undertaking the following process:
We have a couple of different steps for them within our protocol where they essentially run a query to try to identify any ads that our normal models couldn‘t catch. They also do a little bit of research on who the current politicians are that are running in the Washington
State, and they take a look at their pages to see if there‘s any ads that may be running that our machines didn‘t catch, things like that. Q: And you said that these sweeps are done by people?
A: Yes. They‘re done by employees of Meta.
Q: And do the employees have the discretion to craft queries?
A: They do, if they have that capability. It‘s not a necessity for them to be able to craft a query, but if -- we have queries that we created for them that they can use to take a look at, but they‘re investigators. They have the ability to track things down, and they know how to use our internal tooling.
Q: What is your internal tooling? . . .
A: We have various tools that the business integrity team uses to understand what kind of ads people have running, what kind of advertiser they are, what violations they‘ve experienced in the past, the ability to run queries . . . .
(Emphasis added).
In short, an unknown number of persons on a team do a “little bit of research” and may or may not craft queries or use other tools, on a bi-weekly to monthly basis, to catch ads the algorithm may not catch.
As to the argument regarding the algorithm, Meta offered the testimony of an expert witness, Dr. Steven Weber, professor at the School of Information at the University of California Berkeley, who opined that a “machine learning system cannot serve as the exclusive solution to classify whether an advertisement falls under the Washington Disclosure Law” because:
Algorithms are primarily written by people, and a significant amount of coding work and system development is required to write and refine algorithms that are efficient and accurate. In addition to the algorithm, machine learning systems in settings of this kind depend upon “training data.” A set of training data typically consists of manually labeled examples of the classification that the system is
being asked to perform.
(Emphasis added). Dr. Weber elaborated that:
the classification required by the Washington Disclosure Law is a difficult one, even for humans to implement. It is not clear how a human would create such labeled data with a consistently high degree of accuracy, given that whether an ad is “political” is a complex determination that often requires input from multiple individuals. Moreover, even attempting to create this set would take a large amount of time (for example, to have multiple humans review and debate the instances that are challenging to classify) with no guarantee of it being successful or stable.
Creating a sufficiently robust set of training data would be challenging for an additional reason. The need to identify Washington elections, candidates, and ballot propositions in particular means that very specific examples would have to be used in the training data in order to teach the machine to distinguish Washington Political Ads from other types of political advertising. But a candidate or ballot proposition in the State of Washington may only have a handful of ads, making it unlikely for there to be a sufficiently large corpus of training data.
(Emphasis added and omitted.) Further, Dr. Weber contrasted “ads about ‘social issues, elections, or politics,’ was a broader category than that of “ads about Washington state and local elections.” Thus, it would be ”much easier to create a high-quality training data set for these ads” because “there are far more examples available of ads that relate to ‘social issues, elections, or politics.‘” (Emphasis added).
In short, Meta‘s position is that it would take an unspecified but “significant amount of coding work and system development,” followed by an equally unspecified “large amount of” human debate time to address this “complex” definitional challenge.
Finally, Meta offered testimony from Ani Vanesyan, a manager of business
Even so, Ms. Vanesyan in the same deposition admitted modifying the Ad Library would still be “technically feasible.”
The way that the tool is written on the back end in terms of code is done in a particular way. Currently, it‘s done at the country level with all the fields. To create a whole new flow, while it could be technically feasible given some amount of time and certain number of engineers, et cetera, it would include substantial and significant amount of work to make this.
It could also challenge the stability of the tool, because the more complex the code becomes on the back end, the more unstable the tool is. It would also entail quite a few new flows being created to differentiate ads, Washington State versus the rest of the advertising. So this would be nontrivial, and the time spent on this will also potentially be time away from more scalable solutions from extending political authenticity and transparency globally, et cetera.
Q: So it would be costly; is that fair?
A: Not just costly in the sense of, you know, time and resources. I think this would be complex. Again, it would challenge the stability of the tool, which doesn‘t really deal with cost but could potentially jeopardize our commitments of authenticity and transparency if the tool was, let‘s say, frequently not working as well as potentially delay our plans to provide these solutions to more countries.
(Emphasis added).
In sum, as to the burden imposed, Meta asserts that the disclosure law contains several that are unnecessary because revisions to the algorithm would take a “significant amount of work” and the human review would be “complex” and “costly” in “time and resources,” even as it acknowledges that the former is
Although Meta has put forth evidence that complying with the law would be unappealing in some sense, it has failed to establish a genuine issue of material fact because it has not provided evidence that supports an “understanding of the extent to which the burdens are unnecessary, and that requires narrow tailoring.” Bonta, 594 U.S. at 611.
Specifically, Meta‘s argument fails to quantify how “significant” the revisions to the algorithm would be, how much more “time and resources” it would take for human review, or even how much any change in its practices would cost. After years of discovery, we know nothing more about the burden on Meta than high level, even “speculative,” generalities which lack the “specificity” to create a genuine issue of material fact. Bucci, 197 Wn. App. at 326. Although Meta provided evidence about two steps it would need to take, and alludes that it would need to hire multiple employees to, among other things, identify ads for retention and disclosure, nowhere does it identify the specific cost or number of employees required to do so.
Moreover, none of Meta‘s cited evidence indicates that it cannot comply; rather, the evidence only shows, as the State phrases it, that it would be inconsistent with Meta‘s corporate priorities to do so. Meta does not contest, or offer competing evidence to counter, this point. As the superior court found, “[i]n essence, the only reason why Meta refuses to comply with the law is, to put it colloquially, they don‘t want to let the public see how the sausage is made.”
Finally, the question is whether the burdens, however steep, are unnecessary and require narrow tailoring. Bonta, 594 U.S. at 611. Meta‘s argument on these issues addresses none of the provisions of the disclosure law in question, let alone shows how any particular provision may create an unnecessary or untailored burden and thus the argument misses the mark at a more fundamental level.
For Meta to survive summary judgment, it must “set[] forth specific facts which sufficiently rebut the moving party‘s contentions” and not mere “argumentative assertions.” Ranger Ins. Co. v. Pierce Cnty., 164 Wn.2d 545, 552, 192 P.3d 886 (2008) (quoting Meyer v. Univ. of Wash., 105 Wn.2d 847, 852, 719 P.2d 98 (1986)). This first argument fails in that regard.
This conclusion leads us to Meta‘s second argument, which contends at a more granular level that the disclosure requirements of
In response to Meta‘s assertion that requiring the disclosure of information such as the advertiser‘s payment method is unnecessary, the State expands on its argument that such information is necessary to help voters understand who pays for advertising to persuade them. Dr. McGregor explained, “if we are talking about digital political ads, to understand it . . . someone would need to know the source; the cost; the content of the ad; and the targeting parameters.” This information is necessary, Dr. McGregor says, “[b]ecause giving only one part of that and not the other part at the same time could lead to, you know, a deep misinterpretation of what the nature of advertising that was actually happening on the platform was.”
For example, as explained by another State witness, Dr. Travis N. Ridout,
when there is misinformation contained within political ads, I think having that targeting information can help to diagnose that and see whom that information is aimed at.24
We agree with the State that gathering and providing the advertiser‘s identity, intended audience, and the source of their funding, for example, is closely tailored to the State‘s goal of helping a voter to evaluate why they received a certain ad. It is true that a particular piece of the data required by the disclosure law in a given situation may not always align perfectly with the State‘s goals; perhaps the payment method may not be that illuminative to understand a given ad. However, the “[t]ouchstone for exacting scrutiny [under the First Amendment] is whether there is a ‘fit [between the chosen means and the purported ends] that is not necessarily perfect, but reasonable.‘” McManus, 944 F.3d at 523 (emphasis
As to Meta‘s claim that non-Washington residents may request this data for any reason, the State again avers that oftentimes journalists, academics, activists, and others who may not reside in Washington are in the best position to disseminate information about ad data. In many cases, as the State asserts in its brief, “[t]he public relies on third parties to become informed,” including academic researchers, in this case, testifying on behalf of the State. Meta does not dispute this claim. Thus, we agree with the State that Meta does not establish a genuine issue of material fact that the “fit” between who can request information, why they may do so, and the State‘s goal is unnecessarily burdensome. McCutcheon, 572 U.S. at 218.
Meta next contends that it is overly burdensome to have to comply with information requests within two business days,25 but the Supreme Court in McConnell considered a requirement for timely disclosure of advertiser data and concluded that, “[g]iven the relatively short timeframes in which electioneering
In response to Meta‘s contention that requiring it to respond to disclosure requests year-round is overly burdensome, the State argues that, in fact, advertisers purchase ads for political purposes outside of the 21-day period before an election, which is a standard laid out in other disclosure laws. And, as noted, the State has shown that Meta already collects this information in its usual course of business throughout the year. Even if there may be time periods where fewer political ads are posted, we agree with the State that Meta has not established a genuine issue of material fact that a year-round requirement is insufficiently fitted to the government‘s purposes or otherwise unnecessarily burdensome. McCutcheon, 572 U.S. at 218.
Meta also avers that a five-year retention period is overly long, but fails to offer a reason, let alone evidence, to show why retaining ad data for five years is burdensome. Moreover, the sworn testimony of Ms. Vanesyan establishes that Meta‘s Ad Library retains this targeting information as part of its regular business for “at least” seven years:
Q: So the targeting information that is selected by the advertiser, that is not in the Ad Library correct?
A: Yes, that‘s correct.
A: My understanding at this time and recollection is that it‘s also not in the Ad Library report.
Q: But that information is retained by [Meta] for seven years; is that correct?
A: At least seven years, yes.
Meta, thus, does not establish a genuine issue of material fact as to whether the retention period required under the disclosure law is unnecessarily burdensome.
Finally, as to Meta‘s claim that few people have actually requested this data, there is no authority requiring that there be a minimum number of requestors before any given person or the PDC itself avail itself of the disclosure law. See
We conclude that Meta does not prevail on even its more granular arguments either because it has failed to demonstrate any genuine issue of material fact that certain specific provisions of the disclosure law are not narrowly tailored or are unnecessarily burdensome, or because controlling authority does not support these various positions.
As to its third category of arguments, Meta asserts that it is burdened because it faces severe penalties if it fails to comply with the disclosure law.26 In
We conclude that Meta has not met its burden “to present evidence that an issue of material fact remains” as to whether the disclosure law is insufficiently narrowly tailored to its stated purpose. Haley, 25 Wn. App. 2d at 216. In turn, the trial court did not err in concluding the disclosure law withstood the second step of the exacting scrutiny analysis, at least with respect to Meta‘s own claimed burdens.
ii. Standing to Raise Burden on the Rights of Others
Meta next argues that the disclosure law unconstitutionally burdens not only its own First Amendment rights, but also those or other rights of at least two other classes of persons; namely, the privacy interests of the purchaser and viewers of ads, and the free speech rights of upstart, grassroots candidates.
It is axiomatic that “[e]very action shall be prosecuted in the name of the real party in interest.”
Focusing only on the third element of prudential third party standing, Meta fails to establish that, nor can we envision how, there is some “hindrance to the third party‘s ability to protect their own interests.” Meta does not present any evidence or offer any authority indicating that ad purchasers, viewers, or political campaigns cannot litigate their own First Amendment claims. Indeed, political
As to the non-constitutional third-party privacy rights allegedly violated here, it is clear that Meta has no third party standing to raise its constitutional claims based on the right to privacy of another. See, e.g., State v. Farmer, 116 Wn.2d 414, 805 P.2d 200 (1991) (denying standing where a defendant convicted of sexual exploitation of a minor and patronizing a juvenile sex worker challenged the governing criminal statute for violating the victim‘s First Amendment); Keepers, Inc. v. City of Milford, 807 F.3d 24 (2d Cir. 2015) (same).
Accordingly, we decline to consider Meta‘s arguments regarding the rights, constitutional or otherwise, of third parties.
4. McManus Does Not Control
The authority Meta most relies on is McManus, which it claims “explains why the [] Disclosure Law is unconstitutional.” We conclude that McManus is distinguishable.
In McManus, the Fourth Circuit struck down, stated generally, “a Maryland law [that required] newspapers, among other platforms, to publish on their websites, as well as retain for state inspection, certain information about the
Several news outlets sued to enjoin the Maryland law from becoming effective because its “publication and inspection requirements violated the First Amendment-both facially, as a regulation that targets neutral third-party platforms, and also as-applied, as a law encompassing news outlets.” Id. 512.
The court concluded that the Maryland act‘s disclosure and publication requirements were “compelled speech.” Id. 514. The court held that a core problem with Maryland‘s law was that it made “certain political speech more expensive to host than other speech because compliance costs attach to the former and not to the latter.” Id. 516.
Meta argues in its opening brief that “McManus‘s logic applies with ‘even greater force’ here because the Disclosure Law places heavier burdens on platforms and thus is more likely to result in closing off avenues for political speech.” In reply, Meta asserts that the holding in McManus does not turn on the
The Court in McManus decided the case at least in part because “the [Maryland] Act forces news outlets to publish certain information on their websites and, if they fail to do so, empowers the state to seek a court order to have content pulled from these platforms. This brings with it yet another First Amendment infirmity.” McManus, 944 F.3d at 517-18 (emphasis added).
Meta differs significantly from a newspaper. For purposes of the service in question, Meta is comparable to a billboard where advertisers pay, not only to display their messages, but for Meta to decide where to place the billboard, and thus target a particular group of viewers. There is nothing in the record that demonstrates that Meta exercises editorial control over the content, as a newspaper does. Unlike billboards, “[t]he [Supreme] Court has made clear that news products are of a part with parades, political leaflets, fundraising pitches, or similar expressive endeavors.” Id. 520 (emphasis added).
Moreover, the court in McManus repeatedly emphasized that “the core problem with this provision of the Act is that it lacks any readily discernable limits on the ability of government to supervise the operations of the newsroom.” Id. 518-19 (emphasis added). “Not only does [the Maryland act] compel the Publishers to turn over information to state regulators, it also brings the state into an unhealthy entanglement with news outlets,” by requiring that they alter the content of a news product. Id. 518-19. The court held, “[t]he Supreme Court has made clear that when the government tries to interfere with the content of a newspaper or the message of a news outlet, the constitutional difficulties mount.” Id. 517. No such editorial entanglement occurs here with Meta.
Finally, in McManus, one of the requirements of the Maryland law was that newspapers display information to the public at large in their product, id. 514, while the Washington laws only require Meta to disclose the required data to individual requesters in a variety of formats. RCW 42.17A.345; WAC 390-18-050(4)(a)-(b). The speech at issue here, to the extent there is any, is simply transmittal of data Meta already keeps in a form separate and apart from its product platform, and, thus, is not “compelled” in the same way.
In sum, we conclude McManus is distinguishable because it arises from a materially different type of legal framework within a different state statutory scheme as applied to a different type of entity. McManus was not presented with some of the differentiating facts that appear in the case before us, where the requirement (a) is one of only record-keeping and disclosure or inspection of data Meta is already keeping, (b) does not apply to newspapers and their unique and complex constitutional rights, and (c) does not involve the government in editorial decisions, as there are none.
Thus, Meta has not carried its burden, as the nonmoving party on summary judgment, “to create an issue” of material fact. Rice, 167 Wn. App. at 89. Ultimately, Meta‘s response is simply an “argumentative assertion[] that unresolved factual issues remain.” Bucci, 197 Wn. App. at 326 (quoting Ranger Ins. Co., 164 Wn.2d at 552).
B. Express Federal Preemption
A given federal law may avail itself of this well-known principle and preempt a state law expressly or implicitly. Kissan Berry Farm v. Whatcom Farmers Coop., 23 Wn. App. 2d 490, 500, 516 P.3d 821 (2022). Namely, “[i]f ‘Congress expressly withdraws specified powers from a state through a statutory provision,’ there is express preemption.” City of Woodinville v. Eastside Cmty. Rail, LLC, 22 Wn. App. 2d 121, 127, 510 P.3d 355 (2022) (quoting Beatty v. Fish & Wildlife Comm‘n, 185 Wn. App. 426, 454, 341 P.3d 291 (2015)).
Courts gauge a federal law‘s preemptive effect by “‘focus[ing] on the plain wording of [a given federal statute‘s preemption] clause, which necessarily contains the best evidence of Congress’ pre-emptive intent.‘” Commonwealth of Puerto Rico v. Franklin Cal. Tax-free Tr., 579 U.S. 115, 125, 136 S. Ct. 1938, 195 L. Ed. 2d 298 (2016) (quoting Chamber of Commerce of U.S. v. Whiting, 563 U.S. 582, 594, 131 S. Ct. 1968, 179 L. Ed. 2d 1031 (2011)).
“‘We must interpret an express preemption clause narrowly but fairly.‘” Woodinville, 22 Wn. App. 2d at 127 (quoting Kitsap County v. Kitsap Rifle & Revolver Club, 1 Wn. App. 2d 393, 404, 405 P.3d 1026 (2017)). That is, express
Consistent with this authority, our Supreme Court has held that there is a strong presumption against preemption. Nw. Wholesale, Inc. v. Pac Organic Fruit, LLC, 184 Wn.2d 176, 184, 357 P.3d 650 (2015) (“‘any consideration of preemption issues starts with the assumption that the historic police powers of the States are not to be superseded by ... Federal Act unless that is the clear and manifest purpose of Congress.‘“) (quoting Hue v. Farmboy Spray Co., Inc., 127 Wn.2d 67, 78, 896 P.2d 682 (1995)) (internal quotation marks and alterations omitted).
“The meaning of the presumption against preemption is twofold.” Hue, 127 Wn.2d at 79. “First, where Congress expressly addresses state authority in a federal law, the preemptive scope of the federal law should not be extended any further by resort to an implied preemption analysis.” Id. Second, “‘Congress’ enactment of a provision defining the pre-emptive reach of a statute implies that matters beyond that reach are not preempted.‘” Id. (quoting Cipollone v. Liggett Group, Inc., 505 U.S. 504, 517, 112 S. Ct. 2608, 120 L. Ed. 2d 407 (1992)).
“Preemption is a question of law we review de novo.” McKee v. AT&T Corp., 164 Wn.2d 372, 387, 191 P.3d 845 (2008).
1. Section 230 of the Communications Decency Act
“Section 230 was prompted by a state court case holding Prodigy [an online platform] responsible for a libelous message posted on one of its financial message boards.” Fair Hous. Council of San Fernando Valley v. Roommates.com, LLC, 521 F.3d 1157, 1163 (9th Cir. 2008) (discussing Stratton Oakmont, Inc. v. Prodigy Servs. Co., No. 31063/94, 1995 WL 323710, 1995 N.Y. Misc. LEXIS 229 (Sup. Ct. May 24, 1995) (unpublished)). The Roommates.com court observed that, in Stratton Oakmont, “Prodigy had become a ‘publisher’ under state law because it voluntarily deleted some messages from its message boards ‘on the basis of offensiveness and bad taste’ and therefore was legally responsible for the content of defamatory messages that it failed to delete.” Id. (emphasis added) (internal quotation marks omitted) (citing Stratton Oakmont, 1995 WL 323710 at *4). In that situation, “[u]nder the reasoning of Stratton Oakmont, online service providers that voluntarily filter some messages become liable for all messages transmitted, whereas providers that bury their heads in the sand and ignore problematic posts altogether escape liability.” Id. “In passing section 230, Congress sought to spare interactive computer services this grim choice.” Id.
More broadly, Section 230 preemption seeks “two parallel goals[:] . . . ‘to
Section 230 preemption intuitively applies to actions sounding in tort, such as defamation, but courts have made clear that its scope may apply to all civil cases as “many causes of action might be premised on the publication or speaking of what one might call ‘information content.‘” Id. at 1101 (quoting 47 U.S.C. § 230(c)(1)). “Thus, what matters is not the name of the cause of action . . . what matters is whether a cause of action inherently requires the court to treat the defendant as the ‘publisher or speaker’ of content provided by another.” Id. at 1101-02 (quoting 47 U.S.C. § 230(c)(1)).
The Ninth Circuit created a three-part Section 230 preemption test in Barnes. Id. Under this test, Section 230 “only protects from liability (1) a provider or user of an interactive computer service (2) whom a plaintiff seeks to treat, under a state law cause of action, as a publisher or speaker (3) of information provided by another information content provider.” Id. (emphasis added).
In the present appeal, the parties only contest the second prong of Barnes, as there is no dispute that Meta is an “interactive computer service” and that third parties generate the “information content” of the political ads.
By way of a brief taxonomy, courts have held that a plaintiff treats a defendant as a publisher, and thus preemption applies, in two relevant scenarios
Second, a state law is preempted when it “would necessarily require an internet company to monitor third-party content.” HomeAway.com, Inc. v. City of Santa Monica, 918 F.3d 676, 682 (9th Cir. 2019) (emphasis added). HomeAway concerned a municipal ordinance prohibiting short-term rental platforms, such as Airbnb and HomeAway, from processing rental transactions for properties not listed in an official city registry of such properties. Id. at 680. HomeAway argued Section 230 preempted the ordinance because “it implicitly require[d] them ‘to monitor the content of a third-party listing and compare it against the City‘s short term rental registry before allowing any booking to proceed.” Id. at 682.
The court in HomeAway upheld the city ordinance, reasoning that the “booking transaction-content that, while resulting from the third-party listings, is distinct, internal, and nonpublic.” Id. at 683. “While keeping track of the city‘s registry is ‘monitoring’ third-party content in the most basic sense, such conduct cannot be fairly classified as ‘publication’ of third-party content” as the platforms “have no editorial control over the registry whatsoever.” Id. at 683 (some emphasis added).
The court also held that Section 230 does not “suggest that [its] immunity attaches any time a legal duty might lead a company to respond with monitoring
Moreover, a “‘service provider remains liable for its own speech’ and for its own unlawful conduct.” Airbnb, Inc. v. City of Boston, 386 F. Supp. 3d 113, 119 (D. Mass. 2019) (emphasis added) (quoting Universal Commc‘n Sys., Inc. v. Lycos, Inc., 478 F.3d 413, 419 (1st Cir. 2007)). In contrast, liability does not attach for “publication,” which “involves reviewing, editing, and deciding whether to publish or to withdraw from publication third-party content.” Barnes, 570 F.3d at 1102 (emphasis added). As such, if a “provision is aimed at regulating [the defendant‘s] own conduct, and not punishing it for content provided by a third party, [Section 230] does not preempt it.” Airbnb, Inc., 386 F. Supp. 3d at 120 (emphasis added).
In short, a law is preempted under Section 230 if it “(1) makes the defendant liable for publishing certain information to third parties, and (2) seeks to impose liability based on that information‘s improper content.” Henderson v. Source for Public Data, L.P., 53 F. 4th 110, 120-21 (4th Cir. 2022) (emphasis added).
2. Application of Section 230 to Meta
Meta urges us to hold that Section 230 expressly preempts Washington‘s disclosure law. Specifically, Meta argues the disclosure law is “preempted twice over because a defendant‘s being a publisher and engaging in publication are both
Meta fundamentally mischaracterizes the two main aims of
Here, as explained in more detail above, entities covered by the disclosure law are obligated to “maintain current books of account and related materials as provided by rule.”
The disclosure law does not create liability for the actual publication of ads or, crucially, the actual content of an advertisement. Instead, the disclosure law imposes liability squarely for Meta‘s own conduct, namely whether it “maintain[ed] current books of account” that can be “open for public inspection.”
Meta next argues the disclosure law improperly treats Meta as a publisher by “requir[ing] the platform to review third-party content posted on its site.” Specifically, Meta avers that the disclosure law‘s “nuanced definition of ‘political advertising‘” necessarily requires Meta to monitor third-party content in violation of HomeAway.
Even if we were to assume the disclosure law requires Meta to review ads, any more than it already does voluntarily, it “does not proscribe, mandate, or even discuss the content of the” ads beyond identifying the broad, high-level categories of “political advertising or electioneering communications during the election campaign.” HomeAway, 918 F.3d at 683 (emphasis added);
At most, the disclosure law merely tasks Meta with placing ads into “political” and “non-political” categories to facilitate future inspection or disclosure, which it has attempted to do for several years anyway through its Ad Library. The disclosure law does not require Meta to “publish, withdraw, postpone or alter content” contained within those groupings. Schneider, 108 Wn. App. at 463 (quoting Zeran, 129 F.3d at 330) (emphasis added). In other words, while categorization may be monitoring “in the most basic sense, such conduct cannot be fairly classified as ‘publication’ of third-party content” as Meta has “no editorial control over” the ads “whatsoever.” HomeAway, 918 F.3d at 682-83. As such, Meta “face[s] no liability for the content” of political ads. Id. at 684 (emphasis added). In turn, the second prong of the three-part test in Barnes is not met.
The prior analysis resolves the question as a matter of law. However, as a practical matter, it is noteworthy that the disclosure law defines the terms “political advertising or electioneering communications” in a way that simplifies even such minimal monitoring. The definitions target advertisements which “appeal directly or indirectly for . . . support or opposition in any election campaign” or “[c]learly identifies a candidate for a state, local, or judicial office.”
Such ads are readily identifiable because the Washington Secretary of
In HomeAway, platforms merely had to ensure they were not “completing any booking transaction for properties not licensed and listed on the City‘s registry.” Id. at 680. Similarly, Meta can compare the ads against Washington‘s publicly available list of elections and candidates, and then preserve and prepare the information related to such ads for inspection and possible disclosure. In neither case does such “monitoring” constitute “publication” under Section 230.
For these reasons, we hold that Section 230 does not preempt the
Thus, Meta‘s two substantive challenges to the disclosure law have failed and the court did not err in finding the State was “entitled to a judgment as a matter of law.”
Next, we turn to whether the superior court properly assessed the penalties and fees it did.
C. Penalties
The penalty provision of the disclosure law provides that a “person who violates any of the provisions of this chapter may be subject to a civil penalty of not more than ten thousand dollars for each violation.”
Meta challenges the superior court‘s $24,660,000.00 civil penalty for 822 violations of the disclosure law. Meta substantively presents three arguments. First, it argues the number of violations should have been calculated on a per-request basis instead of a per-ad basis. Second, it argues the court improperly imposed the statutory maximum penalty of $10,000 for each violation. Third, it argues the court improperly trebled damages after finding Meta‘s violations were
1. Per-Ad or Per-Request
The disclosure law permits a court to impose a penalty “for each violation.”
At the summary judgment hearing, the superior court initially stated Meta‘s violations would be based on the number of requests. The judge explained that:
enforcement penalties and so on - have to be limited to the requests . . . when somebody makes a request and they don‘t produce it, then they‘re not complying with the law. But before . . . I don‘t think that‘s something I can enforce against them with - absent somebody making a request.
However, the court‘s subsequent written order based Meta‘s violations on the number of undisclosed advertisements. The court did not elaborate as to why it changed its method of calculation, but the effect was to increase the number of violations from 12 to 822. As such, the total penalty increased from $120,000 to $8,220,000.00 which the superior court then trebled to $24,660,000.00.
Bittner provides a helpful analytical framework for examining two competing penalty calculation methods in that it systemically examined (a) the text and legislative history of the statute, (b) the “contextual clues,” such as government documents, which reveal the “the consistency of an agency‘s views,” and (c) the rule of lenity. Bittner, 598 U.S. at 101. We address each in turn.
a. Text and Legislative History
“When interpreting a statute, the court‘s fundamental objective is to ascertain and give effect to the legislature‘s intent. We begin with the plain meaning of the statute. . . . If the meaning of the statute is plain on its face, then we must give effect to that meaning as an expression of legislative intent.” Leander v. Dep‘t. of Ret. Sys., 186 Wn.2d 393, 405, 377 P.3d 199 (2016) (citations omitted); Belleau Woods II, LLC v. City of Bellingham, 150 Wn. App. 228, 240, 208 P.3d 5 (2009) (in gauging a statute‘s plain meaning, “it is appropriate to look at the language in the context of the statutory scheme as a whole.“). We review questions of statutory interpretation de novo. Utter ex rel. State v. Bldg. Indus. Ass‘n of Wash., 182 Wn.2d 398, 406, 341 P.3d 953 (2015).
Meta’s argument is belied first by the FCPA and disclosure law’s overall plain intent to ensure full disclosure of certain records, namely, “contributions and expenditures” of advertisements on its social media platforms. Specifically, the disclosure law states in its introductory provision that “political campaign and lobbying contributions and expenditures be fully disclosed to the public and that secrecy is to be avoided.”
The plain language of the disclosure law evidences its intent to preserve and sunlight relevant data of each individual ad on Meta’s social media platforms, without any requirement that any particular voter or the government step into the sunlight to inspect the records. In this way, Meta’s discrete serial obligation, to preserve and prepare for inspection or “full disclosure” every individual “contribution[] and expenditure[],” is distinct from any individual request for such disclosure and is the proper metric for violations of this tool of transparency.
Further, we may not ignore the legislature’s twice repeated command that the disclosure law “be liberally construed . . . so as to assure continuing public confidence of fairness of elections and governmental processes, and so as to assure that the public interest will be fully protected.”
Indeed, our courts, in a variety of different but related settings, have reiterated the Washington’s campaign finance laws’ intent to ensure full disclosure of individual records. See, e.g., State v. Econ. Dev. Bd. for Tacoma-Pierce County, 9 Wn. App. 2d 1, 11, 441 P.3d 1269 (2019) (quoting Evergreen Freedom Found., 192 Wn.2d at 790) (“The FCPA is designed, in part, ‘to provide the public with full disclosure of information about who funds initiative campaigns and who seeks to influence the initiative process.’”) (quoting LAWS OF 1973, ch. 1); see also State v. Eyman, 24 Wn. App. 2d 795, 815, 521 P.3d 265 (2022) (quoting State v. Grocery Mfrs. Ass’n, 198 Wn.2d 888, 892, 502 P.3d 806 (2022) (Grocery Mfrs. Assn. II)) (“‘The FCPA is an attempt to make elections and politics as fair and transparent as possible.’”). These statements of intent do not rely on the voter prompting or availing themselves of such transparency.
Returning to Bittner, there, the Bank Secrecy Act “[did] not speak of accounts or their number. The word ‘account’ does not even appear. Instead, the relevant legal duty is the duty to file reports.” 598 U.S. at 93. That is, the Bank Secrecy Act’s “statutory obligation is binary. Either one files a report ‘in the way and to the extent the Secretary prescribes,’ or one does not.” Id.
In contrast, while the penalty provision in the disclosure law does not explicitly reference “political advertising” and “electioneering communications[,]” related provisions of the disclosure law and its accompanying regulations expressly reference both units of measurement multiple times. See
Finally, even if we were to hold that the legislature intended that a violation should be calculated on a per-request basis, a requestor would need only file a separate request for each individual advertisement, which would easily circumvent Meta’s proposed legislative intent. We do not interpret statutes to permit such absurd results. Jespersen v. Clark County, 199 Wn. App. 568, 578, 399 P.3d 1209 (2017) (courts must “construe a statute to avoid absurd results.”).
The plain language of
b. “Contextual Clue”
Meta next argues that one “contextual clue” provides support for its claim that the proper measure of damages is to calculate damages per unfulfilled request. Specifically, Meta cites to a document which summarized the PDC’s initial findings as follows: “[Meta] failed to respond to two requests for information, violating
Even assuming arguendo that this document represents the official position of the PDC, “the government’s guidance documents do not control our analysis and cannot displace our independent obligation to interpret the law,” which we conducted de novo above. Bittner, 598 U.S. at 97; Utter, 182 Wn.2d at 406.
Moreover, this statement does not undermine the consistency of the PDC
Meta’s asserted single contextual clue does not weigh in its favor under Bittner’s three-part framework
c. Rule of Lenity
Meta argues that, at a minimum, the disclosure law is ambiguous, requiring an application of the rule of lenity.
Under the rule of lenity, “[i]f after applying rules of statutory construction we conclude that a statute is ambiguous, ‘the rule of lenity requires us to interpret the statute in favor of the defendant absent legislative intent to the contrary.’” State v. Baker, 194 Wn. App. 678, 684, 378 P.3d 243 (2016) (emphasis added) (quoting City of Seattle v. Winebrenner, 167 Wn.2d 451, 462, 219 P.3d 686 (2009)). The
Here, the disclosure law is clearly a civil penal statute as it authorizes “sanctions” and “penalties” for violations.
Therefore, under the Bittner framework, the proper method of calculation is, as a matter of statutory interpretation, by advertisement.33
2. Statutory Maximum Penalties
Under the disclosure law, the statutory maximum penalty is ten thousand
“We review the trial court’s assessment of civil penalties within the statutory limits for an abuse of discretion.” State v. Mandatory Poster Agency, Inc., 199 Wn. App. 506, 525, 398 P.3d 1271 (2017). A court abuses its discretion if “its decision is manifestly unreasonable, or exercised on untenable grounds or for untenable reasons.” Gildon v. Simon Prop. Grp., Inc., 158 Wn.2d 483, 494, 145 P.3d 1196 (2006). “An abuse of discretion is found if the trial court relies on unsupported facts, takes a view that no reasonable person would take, applies the wrong legal standard, or bases its ruling on an erroneous view of the law.” Id.
On appeal, Meta faults the superior court for “not credit[ing] a single one of those factors” enumerated in
- First, Meta has promoted political transparency in Washington, sometimes exceeding Washington’s requirements. . . .
- Second, Meta undisputedly disclosed much of the required information immediately on request. . . .
- Third, there is no dispute that for some of the requests at issue, Meta produced two and sometimes three additional categories of required information. . . .
Fourth, Meta established a formal process to efficiently respond to requests under the [disclosure law]. . . . - Fifth, it is undisputed that none of the requests at issue in this case were made for the core purposes that animate [the disclosure law] —informing voters and ensuring election integrity. . . .
- Sixth, during the PDC’s investigation of Mr. Sanders’s and Mr. Trask’s 2019 complaints, it is undisputed that Meta extensively consulted and cooperated with PDC Staff. . . .
- Seventh, Meta never intentionally or knowingly skewed Washington’s political process by favoring one candidate or cause over another. . . .
- Eighth, penalties in similar cases are significantly lower than the Superior Court’s penalty.34
(Emphasis omitted.)
In its response brief, the State asserts that the following facts justified the court’s decision.
- First, Meta has a history of noncompliance. In 2018, the State filed its first lawsuit for Meta’s failure to have complied with the law since at least 2013, which resulted in a stipulated judgment. . . .
- Second, Meta is the largest social media company in the world, and a significant part of its business model is to sell advertising across its platforms in the United States and in more than 160 countries around the world. . . .
- Third, Meta is extensively involved in campaign finance activity by providing and facilitating the targeting of political ads intended to influence elections in Washington. . . .
- Fourth, Meta did not act in good faith and refused to take responsibility for its violations. Despite stipulating to a judgment in
2018, Meta repeatedly continued to violate the [disclosure law]. . . . - Finally, the superior court correctly recognized Meta’s brazen behavior. Meta continues now to arrogantly justify its pattern of intentional violations by arguing—without any factual support—that its own business model and decisions made compliance “impossible.”
(Emphasis omitted.)
The court explained its reasoning for imposing the $10,000 statutory maximum penalty at length. First, it found Meta “didn’t really make a significant effort to try and comply, either in terms of providing any targeting information, which is clearly required by the law, or in terms of making a serious attempt to comply in a timely fashion.” Second, the court found it “would be very easy for [Meta] to simply organize the information in a way that would make it readily available” as “[it] had to take it all in digitally in the first place in order to get the ad.” Additionally, the court considered “Meta’s compliance history” which included a “pattern of knowing and repeated violations” as well as “Meta’s extensive experience with campaign finance law.” Finally, the court noted “Meta’s lack of good faith and failure to acknowledge and take reasonability for its violations.”
Again, the language of the disclosure law’s penalty provision is highly discretionary.
Ultimately, we review a superior court’s decision to impose a penalty within the statutory limits for an abuse of discretion. Mandatory Poster, 199 Wn. App. at 525. In similar circumstances, we “accept the trial court’s resolution of credibility and conflicting substantial evidence, and its choice of possible reasonable inferences.” Kreidler v. Cascade Nat’l. Ins. Co., 179 Wn. App. 851, 861, 321 P.3d 281 (2014). While another judge may have considered the factors and evidence differently, Meta failed to establish the superior court abused its discretion as it did not show its factual findings were unsupported by the record. Gildon, Inc., 158 Wn.2d at 494.
Thus, we affirm the superior court’s decision to assess the $10,000 statutory maximum penalty for each violation.35
3. Treble Damages
The disclosure law provides that if “the violation is found to have been intentional, the amount of the judgment, which shall for this purpose include the costs, may be trebled as punitive damages.”
Our Supreme Court has held that “[w]hether treble damages should be imposed is thus clearly a discretionary decision for the trial court, which can be made only after the court determines that a violation was intentional.” Grocery Mfrs. Assn. I, 195 Wn.2d at 473. The meaning of “intentional” is not defined in the disclosure law, meaning we must apply its “usual and ordinary meaning” unless context dictates otherwise or the ordinary meaning would lead to absurd results. Id. at 471. Our Supreme Court, in examining the FCPA, has held that “[t]he ordinary meaning of intentional, in both civil and criminal contexts, requires intent to accomplish an unlawful act, but not subjective knowledge that the act is unlawful.” Id.
Meta argues the superior court improperly imposed treble damages and
The superior court did not abuse its discretion by trebling damages. At a minimum, the court’s decision is tenably based on at least the following two intentional acts, each of which flatly contravened unambiguous provisions of the disclosure law.
First, Meta admitted it unilaterally issued a “Request for Information from Facebook, Inc. Regarding Washington Political Advertising” form to requestors, which stated that a request “cannot exceed one year and must begin January 1, 2019 or later.” These limitations plainly violate the disclosure law, which has no such temporal restrictions and, instead, requires that platforms retain the pertinent records “for a period of no less than five years after the date of the applicable election.”
Second, and equally substantively, the superior court found that Meta “intentionally redacted required information from the records it provided in response to requests.” This finding is adequately supported by the record. A Meta employee testified to his personal knowledge that that Meta did not “provide any
The superior court did not abuse its discretion in finding that Meta’s violations were intentional and in awarding treble damages on that basis, as those decisions are supported in the record. Gildon, 158 Wn.2d at 494.37
Accordingly, we hold that the superior court did not err in imposing the penalty in the manner and substance it chose.
D. Attorney Fees
The disclosure law provides that “[i]n any action brought under this chapter, the court may award to the commission all reasonable costs of investigation and trial, including reasonable attorneys’ fees to be fixed by the court.”
In the trial court, the State requested, and was granted, its costs and reasonable attorneys’ fees pursuant to
Thus, we grant the State’s request for its fees and costs on appeal. The State may perfect its request before the commissioners of this court consistent with our rules of appellate procedure.
We affirm.
WE CONCUR:
Díaz, J.