In Justice Thomas’s dissent to the denial of certiorari in Khatibi v. Yip, joined by Justice Alito (Justice Gorsuch voted in favor of granting the petition but did not join the dissent):
In 2022, California instituted a requirement for private instructors in privately operated and privately accredited continuing medical education (CME) courses to address how implicit bias purportedly shapes healthcare outcomes. The Ninth Circuit concluded that First Amendment review did not apply to this compelled-speech mandate, arguing that the broad scope of California’s CME regime makes the course content government speech rather than private speech. That reasoning misconstrues this Court’s precedents and the Constitution. First Amendment protections do not yield simply because a state imposes extensive regulation on speech.
I would have granted certiorari to clarify the government-speech doctrine. At the least, the Ninth Circuit misapplied Shurtleff v. Boston (2022), which directs courts to identify government speech by examining (1) the history of the expression at issue, (2) how the public is likely to view who is speaking—the government or a private person—and (3) the extent to which the government has actively shaped or controlled the expression. The Ninth Circuit collapsed these distinct inquiries by centering on the breadth of the regulatory scheme at each step. More fundamentally, however, Shurtleff is steering courts astray.
I joined Justice ALITO’s concurrence in Shurtleff, which cautioned that the majority’s three-factor test could push courts to conclude that extensive government regulation converts a private message into government speech excluded from First Amendment scrutiny. As the Ninth Circuit’s ruling demonstrates, those concerns were well-founded. This Court should reevaluate whether the Shurtleff test is the correct framework for determining government speech…
Physicians licensed in California must complete 50 hours of approved CME every two years. CME activities eligible for credit are frequently organized by private organizations, certified by private accreditors, and taught by private instructors. The State does not preapprove course content but retains the power to audit courses to ensure compliance with CME standards.
The Medical Board of California “adopts and administers [CME] standards,” with only a few provisions touching course content. Moreover, the State requires that CME courses broadly relate to the medical profession. They must “serve to maintain, develop, or increase the knowledge, skills, and professional performance that a physician and surgeon uses to provide care, or to improve the quality of care provided to patients.” In 2019, however, the California Legislature added a stipulation that after January 1, 2022, all CME courses must “contain curriculum that includes the understanding of implicit bias.” To meet that requirement, CME courses must include “[e]xamples of how implicit bias affects perceptions and treatment decisions of physicians and surgeons, leading to disparities in health outcomes” or “[s]trategies to address how unintended biases in decisionmaking may contribute to health care disparities.”
Petitioners Dr. Azadeh Khatibi and Do No Harm challenge § 2190.1(d) as compelled speech in violation of the First Amendment. Dr. Khatibi, a California-licensed ophthalmologist, designs and teaches CME courses. She “disagrees that implicit bias is the primary factor driving disparities in healthcare” and does not wish to devote class time to that topic. Do No Harm is a nonprofit organization; at least one of its members teaches CME courses and believes that implicit-bias trainings “have not been shown to successfully reduce barriers to healthcare, and instead risk infecting healthcare decisions with divisive and discriminatory ideas.” Yet California’s implicit-bias-training requirement leaves petitioners no option but to discuss the subject in a way that endorses the State’s view that implicit bias causes “disparities in health outcomes….”
The Ninth Circuit held that all three Shurtleff factors point to CME programs eligible for Board credit as constituting government speech. Throughout the analysis, the court stressed the breadth of the State’s regulation. It noted that California “has long regulated the practice of medicine as an exercise of the police power” and has “adopted, updated, and enforced CME standards.” It reasoned that the public would likely view the speaker in a CME course as California’s licensees, since the profession is heavily regulated. It emphasized California’s extensive control over accredited CMEs as a key factor in determining speech governmentally controlled. As Judge VanDyke stressed in his dissent from the denial of rehearing en banc, the panel’s approach effectively boiled down to a single question: whether the government “heavily” or “actively” regulates CMEs? …
By applying the Shurtleff factors in a rigid and blinkered way, the Ninth Circuit lost sight of the central inquiry for government speech: whether the government is speaking for itself or merely regulating private expression.
The government-speech doctrine occupies a narrow niche within our First Amendment jurisprudence. The First Amendment protects the people’s right to free expression from government interference; it does not bar the government from speaking for itself. “But courts must exercise great caution when a government claims that speech by private speakers is really government speech.” “If private speech could be recast as government speech just by attaching a government seal of approval, the government could muffle unpopular viewpoints,” or, as here, compel private speakers to convey government messages without facing constitutional scrutiny.
This Court has repeatedly exercised “great caution” before labeling a message as government speech. In Johanns v. Livestock Marketing Assn (2005), we held that government-sponsored beef advertisements qualified as government speech because the government determined “the overarching message” of the ads, helped craft proposals, reviewed every advertisement “for substance and for wording,” and the Secretary of Agriculture “exercised final approval authority over every word used in every” ad. In Pleasant Grove City v. Summum (2009), we held that the permanent display of monuments in a public park was government speech because the City “effectively controlled the messages sent by the monuments in the Park by exercising final approval authority over their selection.” And in Walker v. Texas Div., Sons of Confederate Veterans, Inc. (2015), we held that the messages on government-issued license plates, over which the State “maintain[ed] direct control,” were government speech.
Since Walker—which we have described as the “outer bound” of the government-speech doctrine, Matal v. Tam (2017)—we have been careful to distinguish between government control over messages and government control over the broader framework through which those messages are communicated. We have held, for instance, that trademarks are private speech despite the government’s substantial involvement in the registration process. And this Court ruled that flags flown outside Boston’s City Hall did not amount to government speech in part because the city allowed private groups to fly their own colors. Across these decisions, the controlling question remained whether the government intends to speak for itself or to regulate private expression.
In plain terms, and viewed against our precedents, California’s attempt to compel private individuals in privately accredited CME courses to conduct implicit-bias training appears to be an effort to regulate private expression. … First, California has not historically used CME to convey government messages. While the State requires CME courses to relate to medical practice and occasionally to cover specific topics, it has not itself dictated course content in the way a state would regulate public messaging.
Second, because California “relies on the participation of private parties in executing the CME scheme,” the public is unlikely to perceive a private CME course as government speech. As the Ninth Circuit acknowledged, Dr. Khatibi plausibly alleged that attendees view her as the person responsible for CME content.
Third, although California regulates the medical field, licenses physicians, and sets CME credit requirements, there is little evidence that California shapes the messages conveyed in CME courses. Despite its extensive regulatory narrative, the Ninth Circuit did not identify any state law that compels private speakers in a CME course to adopt a particular viewpoint—even on contested medical issues. The implicit-bias-training mandate stands out as a notable departure from that practice. It should have been straightforward for the lower courts to conclude that California’s implicit-bias training regulates private speech.
More troubling than the Ninth Circuit’s misapplication of the Shurtleff factors are the consequences of the ruling. Under the Ninth Circuit’s reading of government speech, might California require a CME instructor to relay the state’s preferred stance on abortion access? Could the state compel an instructor to tell participants that gender-transition surgery is ethically required for some patients?
What about other professions subject to California licensing and oversight? At the next statewide legal education gathering, could California insist that a keynote speaker oppose affirmative action in hiring within law firms? If the Ninth Circuit’s logic stands, extensive government regulation of a field could convert a private speaker’s message into government speech that escapes First Amendment protection in any of these scenarios. That result would not align with this Court’s precedents, nor with the Constitution….
If the Ninth Circuit’s ruling truly reflects the kind of outcome that Shurtleff‘s three-factor test would permit, we ought to rethink Shurtleff. Before Shurtleff, this Court treated history, public perception, and government control as “relevant” but not obligatory “considerations” in identifying government speech. When Shurtleff framed these notions as a three-factor test, several Members warned that the approach would be impractical and would obscure the central inquiry in government-speech cases: whether the government is speaking for itself or regulating private speech. “Like any factorized analysis,” the majority’s approach in Shurtleff fails to offer a principled method for deciding cases. Moreover, a narrow focus on history and the scope of government control over speech risks giving courts a false sense that state regulation can turn private speech into government speech.
Within only a few years, those concerns are bearing fruit. The Ninth Circuit leaned on California’s CME regulation to conclude that requiring private speakers to convey a government message was not compelled speech but government speech. That result embodies the paradox foreseen by the Shurtleff concurrence: that government action regulating private speech could now qualify as government speech outside First Amendment constraints. But neither compelled speech nor censorship become constitutional through aggressive and direct application. It cannot be the case that compelled speech becomes government speech simply because the government compels it.
I would grant certiorari to set a correct course. In my view, the essential inquiry in a government-speech case is straightforward: Is the government speaking, or is it regulating private expression? The challenged government activity must amount to government speech in the literal sense—deliberate communication of a government-determined message by someone acting with authority to speak for the government. Moreover, the government must not rely on means that curb the speech of individuals acting in a private capacity, because doing so reveals that the government has shifted from speaker to regulator.
Under this framework, California’s law clearly invites First Amendment scrutiny. The implicit-bias-training obligation is not speech conducted by a government actor; it requires private speakers to convey certain messages. Although the government may appoint private persons as its agents to deliver its message, Dr. Khatibi and Do No Harm’s members do not volunteer to assume a public or quasi-public role enabling them to speak for the government. They share their own messages through the CME courses they develop and administer, and they do so in their private capacities. Even government speech in the literal sense is not shielded from First Amendment challenges when it employs means that restrict private expression, such as compelled speech. When, as here, the government refrains from issuing its own messages and instead compels private speakers to articulate the government’s viewpoint, the First Amendment applies….
A State cannot escape First Amendment scrutiny through regulation. We should not permit the government-speech doctrine to function as a veil for censorship or as a tool for compelled speech.