Chiles v. Salazar and the Limits of Viewpoint Discrimination Doctrine in the Counseling Room

Conversion therapy, defined broadly as any therapeutic practice that attempts to change an individual’s sexual orientation or gender identity toward a heterosexual or cisgender norm, has been condemned by the American Psychological Association, the American Medical Association, and every cognate professional body as both clinically ineffective and affirmatively injurious to patient welfare [1]. More than two dozen states, as well as the District of Columbia, have enacted laws prohibiting licensed healthcare providers from subjecting minors to such practices [2]. That condemnation rests on solid empirical ground, and such legislative response is a legitimate exercise of state power to protect vulnerable minors from harm. The question this article addresses is therefore not whether such harm is real, but whether Colorado’s particular enactment, which affirmatively permitted counseling directed at the affirmation or exploration of gender identity and sexual orientation while categorically forbidding counseling directed at their alteration, could withstand First Amendment scrutiny, given that it predicated liability on the viewpoint a counselor expressed rather than on the nature of the clinical act performed.

In March 2026, the Supreme Court answered that question in the negative, holding 8-1 in Chiles v. Salazar that the statute, as applied to a licensed counselor who used only talk therapy, regulated speech on the basis of viewpoint and must therefore be reviewed under strict scrutiny [3]. The judgment was correct. Colorado’s statutory structure, which permitted counseling aimed at affirming or exploring gender identity and sexual orientation while simultaneously forbidding counseling aimed at changing those attributes, was a textbook case of viewpoint discrimination that no defensible application of the First Amendment could have sustained. Yet the majority opinion by Justice Gorsuch, while reaching the right result, left the hardest questions unanswered. It offered no workable principle for distinguishing regulated professional conduct from regulated speech, and having identified viewpoint discrimination, it treated that label as effectively dispositive without conducting any strict-scrutiny analysis. This article argues that Chiles was rightly decided on the viewpoint discrimination holding, but that its incomplete reasoning and its unaddressed tension with United States v. Skrmetti (2025) leave the law of professional speech in a state of uncertainty that only future opinions can resolve [4].

Colorado’s Statutory Structure

Colorado’s Minor Conversion Therapy Law, enacted in 2019, prohibited licensed healthcare providers from engaging in “any practice or treatment […] that attempts […] to change an individual’s sexual orientation or gender identity,” as well as any “effor[t] to change behaviors or gender expressions or to eliminate or reduce sexual or romantic attraction or feelings toward individuals of the same sex” [5]. Violations could result in fines, probation, or loss of one’s professional license. Crucially, however, Colorado did not remove sexuality and gender identity from the licensed therapeutic context altogether. The statute expressly permitted counseling that provided “[a]cceptance, support, and understanding for the facilitation of an individual’s […] identity exploration and development,” and it equally permitted “[a]ssistance to a person undergoing gender transition” [6]. The law, therefore, drew a line between two competing directions of therapeutic conversation, authorizing one and forbidding the other.

Kaley Chiles, a licensed professional counselor in Colorado Springs, brought the challenge. She offered only talk therapy: no medication, no aversive physical stimuli, no interventions of any kind other than speech [7]. Some of her clients sought help accepting their sexual orientation or pursuing gender transition. Others, she alleged, wanted help to “reduce or eliminate unwanted sexual attractions, change sexual behaviors, or grow in the experience of harmony with [their] bod[ies]” [8]. Colorado law permitted her to counsel the first group toward their goals and forbade her from counseling the second group toward theirs. The only operative variable was the normative direction of the therapeutic outcome the client sought. Of particular significance to the constitutional analysis was a concession Colorado’s Solicitor General made at oral argument: unlicensed life coaches and other non-licensees could “lawfully convey the advice” that Chiles sought to offer, yet she faced professional discipline for the same words [9]. That concession exposed the law’s operative category. If the aim were to protect minors from a harmful viewpoint, the harm would not become constitutionally tolerable when expressed by an unlicensed speaker. The differential treatment of licensed and unlicensed speakers revealed that what Colorado regulated was not a dangerous therapeutic technique but the licensed professional’s expression of a disfavored view.

From Professional Conduct to Viewpoint-Regulated Speech

The lower courts had classified Colorado’s statute as a regulation of professional therapeutic conduct rather than speech, and on that framing applied only rational basis review [10]. That approach tracked the logic of the Ninth Circuit’s decision in Pickup v. Brown (2014), which treated psychotherapy as professional conduct carried out through words, such that those words could be regulated as part of the conduct without triggering standard speech scrutiny [11]. In Chiles, the Court rejected that framing by grounding its analysis in National Institute of Family and Life Advocates v. Becerra (2018), which had already foreclosed the existence of a distinct professional speech category subject to reduced First Amendment protection. As the majority in Chiles stressed, NIFLA had “expressly rejected the notion that professional speech is subject to diminished constitutional protection” [12]. Speech does not lose its constitutional status simply because it is uttered by a licensed professional to a paying patient.

NIFLA also distinguished between two types of laws that purportedly regulate professional speech. Laws that regulate conduct with only incidental effects on speech may survive with lesser scrutiny, while laws that regulate “speech as speech” trigger the standard First Amendment framework [13]. The Court applied that distinction to find the answer in Chiles’s case straightforward. A law does not escape First Amendment review merely because it operates within a licensed profession; the relevant question is whether the law regulates speech “in the case at hand” [14]. As applied to Chiles, the answer was plain: “All she does is speak, and speech is all Colorado seeks to regulate” [15]. Her speech did not become conduct just because the State chose to call it a “treatment” or “therapeutic modality.” As the Court put it, “[t]he First Amendment is no word game,” and the rights it protects “cannot be renamed away or their protections nullified by mere labels” [16]. Colorado’s law was therefore not a regulation of conduct that incidentally burdened speech; it was a direct regulation of speech that selectively permitted expression from one side of a normative dispute while silencing the other. That structure is viewpoint discrimination, an “egregious form” of content regulation, and a form from which governments in this country must “nearly always abstain” [17].

Was Chiles Rightly Decided?

The Court reached the correct decision. The viewpoint problem embedded in Colorado’s statutory structure was not subtle or contested. The law openly distinguished between therapeutic conversations that supported identity exploration and those that sought to change it, effectively “prescrib[ing] what views [Chiles] may and may not express” [18]. Justice Alito observed at oral argument that the law dictated opposite results in those two situations based solely on the viewpoint expressed, adding that it “[l]ooks like blatant viewpoint discrimination” [19]. That characterization was accurate. Colorado’s only available response, that the law regulated treatment methods rather than speech, collapsed once the Court established that there was nothing to Chiles’s practice other than speech. One cannot suppress a treatment method rather than a viewpoint when the treatment method consists entirely of speaking from a particular viewpoint.

The majority was also correct to reject Colorado’s proposed analogies. Licensing, the Court explained, has traditionally addressed professional qualifications; it has not historically “dictated a professional’s point of view” [20]. Informed-consent doctrine requires disclosure of factual, uncontroversial information tied to a specific physical procedure; in Chiles’s case, there was no separate physical procedure, and Colorado was not compelling disclosure but “seek[ing] to silence a viewpoint [Chiles] wish[ed] to express” [21]. Malpractice preserves what the Court aptly called “breathing room for protected speech” by conditioning liability on proof of actual injury, breach of a discernible standard of care, and causation [22]. By contrast, Colorado’s law threatened “fines, probation, and the loss of [Chiles’s] licens[e] simply for expressing a particular view” [23]. None of those structures supports a law that singles out one side of a contested therapeutic question for suppression. The broader principle finds support across a line of cases running from West Virginia State Board of Education v. Barnette (1943), where the Court declared that no official “high or petty” may “command our tongues or silence our voices,” through Rosenberger v. Rector and Visitors of the University of Virginia (1995) and Iancu v. Brunetti (2019), each of which condemned governmental attempts to designate one side of a contested debate as the permissible side [24].

Justice Jackson’s dissent nonetheless raises arguments that deserve more serious engagement than the majority gave them. Jackson contended that Chiles was “not speaking in the ether” but was speaking as a licensed mental-health provider to minor patients in a system purpose-built to distinguish competent care from harmful practice [25]. On her account, the regulation of medical standards is inherently viewpoint-based in application, because “[w]hen a State establishes a standard of care, or punishes a doctor for providing care outside of that standard, it necessarily limits what medical professionals can say and do on the basis of viewpoint” [26]. That is not a defect of Colorado’s law; it is the very nature of standards-based professional regulation. As Jackson put it, drawing on a long line of scholarship, “[i]n the context of medical practice we insist upon competence, not debate” [27]. She drew on Planned Parenthood of Southeastern Pennsylvania v. Casey (1992), in which the Court upheld Pennsylvania’s informed-consent requirements over a First Amendment challenge from physicians, reasoning that those rights were implicated “only as part of the practice of medicine, subject to reasonable licensing and regulation by the State” [28]. Jackson would characterize Colorado’s ban as an exercise of that same regulatory authority.

That is the dissent’s best argument, and the majority dismissed it too quickly. The majority effectively collapsed Casey into the narrower informed-consent doctrine, treating it as a precedent solely about factual disclosures tied to physical procedures. But Casey’s reasoning about medical practice was broader than any single disclosure mechanism; it suggested that the professional relationship and the patient vulnerability it creates justify a regulatory latitude that ordinary public discourse does not. Where the dissent ultimately fails, however, is on the specific statutory structure at issue. Jackson’s general argument succeeds in defending the proposition that states may regulate professional therapeutic conduct; it does not succeed in defending this particular structure. A state may prohibit a therapist from administering aversive physical stimuli or impose procedural safeguards before a therapist engages minors on contested identity questions. What Colorado’s law did, as applied to Chiles, was permit licensed counselors to speak from one side of a normative dispute and prohibit them from speaking from the other. Even Justice Jackson acknowledged that Colorado had “engaged in viewpoint discrimination” [29]. The dissent’s appeal to professional deference provides no answer to that conceded fact, because the viewpoint problem exists regardless of whether the context is professional or not.

The Analytical Shortcut and the Skrmetti Tension

The majority’s central failure was not in identifying viewpoint discrimination but in the sparse guidance its opinion provided for the inquiry that must now proceed on remand. Having determined that strict scrutiny applied, the Court devoted the remainder of its opinion to rejecting Colorado’s proposed analogies and returned the case to the lower courts rather than resolving the ultimate constitutional question itself. The question presented had been whether strict scrutiny governed at all, not whether the law could survive it, and so the remand left the harder analytical work for another day. The majority closed by declaring that “[t]he First Amendment stands as a shield against any effort to enforce orthodoxy in thought or speech” and that laws like Colorado’s “represent an ‘egregious’ assault on both” the right to think freely and the faith in the free marketplace of ideas [31]. Those are powerful commitments, but they are not a substitute for the analysis the strict-scrutiny framework demands. The pattern of declaring victory upon labeling is also visible in 303 Creative LLC v. Elenis (2023), another Gorsuch opinion in which the Court identified a First Amendment violation and resolved it without seriously applying the balancing framework its doctrinal category nominally required. In both cases, the constitutional category does the entire work of the analysis, and the strict-scrutiny test becomes vestigial. This matters in Chiles for a concrete reason. Colorado had genuine evidence of harm; the major professional medical and psychological associations had concluded that conversion therapy causes measurable harm to minors, including heightened rates of depression, anxiety, and suicidal ideation [33]. As the majority itself acknowledged, “[t]he people lose” whenever “the government transforms prevailing opinion into enforced conformity,” but it never asked whether the evidence of harm here was strong enough to constitute just the kind of compelling interest that might satisfy even the most demanding standard [34]. A properly conducted strict-scrutiny analysis might have reached the same invalidating conclusion; the problem is that no such analysis was performed.

More troubling still is the judgment’s unaddressed tension with United States v. Skrmetti (2025), decided just one year earlier. In Skrmetti, the Court upheld Tennessee’s restriction on gender-affirming medical procedures for minors under rational basis review, giving the state considerable deference in making decisions about the appropriate scope of medical treatments for young people [35]. Justice Jackson pointed directly to that case in her dissent, noting that the State may permissibly “prohibi[t] the administration of specific drugs for particular medical uses but not for others” [36]. The tension with Chiles is difficult to dismiss. In Skrmetti, a state’s prohibition on medical procedures available to one class of minors but not another was reviewed deferentially and upheld. In Chiles, a state’s prohibition on therapeutic speech available to one class of minors but not another was treated as presumptively unconstitutional. The Court’s resolution of the differential-treatment problem in the two cases points in precisely opposite directions. One might defend the distinction on the ground that Skrmetti involved medical conduct while Chiles involved speech, and that is a distinction with genuine significance. But if that distinction carries the entire weight of the divergent outcomes, the Court needed to say so and to explain why the conduct-speech boundary was the operative constitutional variable. Chiles said nothing about Skrmetti. That silence is a problem waiting to become a future case.

What the Judgment Left Unresolved

Justice Kagan’s concurrence, joined by Justice Sotomayor, contains the most instructive passage in the entire decision. She agreed that Colorado’s law presented a “textbook” case of viewpoint discrimination because it “distinguishes between two opposed sets of ideas,” namely the one resisting and the other reflecting the State’s own preferred position on how counselors should speak with minors about sexuality and gender identity [37]. She then drew a line the majority did not address, noting that if Colorado had instead enacted a “content-based but viewpoint-neutral” law, that would raise a “different and more difficult question” [38]. That framing, however, is harder to act on, because it assumes a symmetry between the two practices that does not hold at the definitional level. Conversion therapy is characterized precisely by its aim to produce change in a client’s sexual orientation or gender identity. Affirming counseling does not seek to alter those attributes; it supports what the client already experiences. Any restriction keyed to the therapeutic aim of change will therefore fall on only one set of practices, not because a legislature has expressed a preference, but because only one of the two practices is engaged in producing change will therefore fall on only one set of practices not because a legislature has expressed a preference, but because only one of the two practices is engaged in producing change in the first place. The notion of a viewpoint-neutral conversion therapy ban thus risks being incoherent as a matter of definition, and any genuinely neutral regulatory alternative would require a different structural approach, perhaps one governing the process by which counselors engage minors on identity questions rather than the permitted direction of the conversation [39]. Whatever form such a law might take, being content-based, it would still attract demanding scrutiny, and whether the evidence of harm from conversion therapy is sufficiently specific and well-established to satisfy that demanding standard remains an open question, and one the Court has yet to answer. Kagan herself acknowledged that the difference between viewpoint-based and viewpoint-neutral content regulation in the healthcare context could prove “decisive,” but deferred fuller consideration of that question for another day [40].

Alternatively, states may pursue regulation through malpractice, fraud, or general professional competency standards. These alternatives differ materially from prospective licensing bans in timing and structure. Malpractice requires proof of injury, causation, and breach of a standard of care, a structure that intervenes only after harm has already been alleged rather than before it occurs [41]. Fraud captures narrower cases, principally those in which a therapist misrepresented the efficacy or safety of the treatment. General competency standards are available in principle but become contested when the alleged substandard practice is purely verbal and is disputed along ideological as well as empirical lines. For vulnerable minors whose parents may be directing them toward conversion-oriented therapists, the difference between prospective prohibition and retrospective liability is the difference between the harmful encounter not happening and an adversarial proceeding initiated after it has taken place.

Regulatory Consequences

The stakes of Chiles are substantial. As of 2026, twenty-three states and the District of Columbia prohibit licensed healthcare providers from subjecting minors to conversion therapy, while four states and one territory impose more limited restrictions [42]. These regimes operated prospectively, prohibiting specified forms of counseling before any individualized showing of harm was required. The legislative judgment behind them was supported by the major mental health and medical professional organizations and by consistent survey data. The Trevor Project’s 2024 survey of more than 18,000 LGBTQ+ young people reported that 39 percent had seriously considered attempting suicide in the preceding year, 12 percent had attempted suicide, and 13 percent had been threatened with or subjected to conversion therapy [43]. Those figures explain why legislatures favored prophylactic licensing rules over mechanisms that required waiting for individualized injury to materialize. They also constitute the compelling interest showing that Colorado must now make on remand, and that showing may yet succeed if the evidentiary record on the harms specifically attributable to talk therapy, as distinct from aversive physical techniques, is sufficient to survive strict scrutiny.

If the compelling-interest showing fails on remand, or if other states decline the risk of litigation and migrate toward less restrictive regulatory mechanisms, the shift will be meaningful. The prophylactic model of protection will give way to retrospective liability, and the timing of legal intervention will shift from before the therapeutic encounter to after it. That change makes the legal response slower. The majority’s failure to engage seriously with the compelling-interest and narrow-tailoring components of strict scrutiny leaves states without clear guidance on what an adequate regulatory scheme would look like. As Justice Jackson cautioned in dissent, the ruling “threatens to impair States’ ability to regulate the provision of medical care in any respect” and risks ushering in an era in which talk therapists “broadly wield[] their newfound constitutional right to provide substandard medical care” [44]. Whether that alarm proves prescient depends on how seriously lower courts engage with the strict-scrutiny inquiry on remand. On the record of Chiles itself, the answer remains, pointedly, open.

Conclusion

Chiles v. Salazar establishes that the counseling room is not immune from the First Amendment’s prohibition against viewpoint regulation. Colorado’s statutory structure, which permitted counseling in one therapeutic direction while prohibiting it in the other, was a genuine case of viewpoint discrimination that the Court was right to condemn. The majority’s analysis is sound, and the result is correct. Yet Chiles is also a decision that left the hard work undone. By declining to conduct any strict-scrutiny analysis after finding viewpoint discrimination, the majority foreclosed neither the possibility that Colorado could sustain its law under a more demanding examination, nor the possibility that a reformulated, viewpoint-neutral ban could survive constitutional review. The unaddressed tension with Skrmetti compounds that uncertainty. A Court that deferred to state authority over minor healthcare decisions in 2025 and restricted that same authority in 2026 owes an account of the principle that distinguishes the two. That account has not been given. As Justice Jackson warned in dissent, to do “anything else opens a dangerous can of worms” [45]. Until the Court provides that account, Chiles v. Salazar will stand as a decision rightly decided in outcome, incompletely reasoned in analysis, and in need of the further elaboration that only future litigation can supply.

[1] American Psychological Association, Report of the American Psychological Association Task Force on Appropriate Therapeutic Responses to Sexual Orientation 2 (2009); American Medical Association, Policy H-160.991, Reparative (Conversion) Therapy (2019); American Academy of Pediatrics, Policy Statement: Ensuring Comprehensive Care and Support for Transgender and Gender-Diverse Children and Adolescents, 142 Pediatrics e20182162 (2018).
[2] Movement Advancement Project, “Conversion ‘Therapy’ Laws,” last updated March 30, 2026, https://www.lgbtmap.org/equality-maps/conversion\_therapy.
[3] Chiles v. Salazar, 607 U.S. ___ (2026). Justice Gorsuch delivered the opinion of the Court; Justice Kagan wrote a concurring opinion joined by Justice Sotomayor; Justice Jackson dissented.
[4] United States v. Skrmetti, 605 U.S. 495 (2025).
[5] Colo. Rev. Stat. §12-245-202(3.5)(a) (2025); Chiles v. Salazar, 607 U.S. ___, slip op. at 3 (2026).
[6] Colo. Rev. Stat. §12-245-202(3.5)(b)(I)-(II) (2025); Chiles v. Salazar, 607 U.S. ___, slip op. at 3 (2026).
[7] Chiles v. Salazar, 607 U.S. ___, slip op. at 1-2 (2026).
[8] Id. at 2.
[9] Transcript of Oral Argument at 58, Chiles v. Salazar, No. 24-539 (Oct. 7, 2025). Colorado counsel conceded at argument that life coaches and other nonlicensees could lawfully convey the same advice Chiles sought to offer to clients.
[10] Chiles v. Salazar, 607 U.S. ___, slip op. at 6 (2026); see also Pickup v. Brown, 740 F.3d 1208, 1227-31 (9th Cir. 2014); King v. Governor of New Jersey, 767 F.3d 216, 232-37 (3d Cir. 2014).
[11] Pickup v. Brown, 740 F.3d 1208, 1227-31 (9th Cir. 2014).
[12] Chiles v. Salazar, 607 U.S. ___, slip op. at 14 (2026) (quoting Nat’l Inst. of Fam. and Life Advocs. v. Becerra, 585 U.S. 755, 767 (2018)).
[13] Nat’l Inst. of Fam. and Life Advocs. v. Becerra, 585 U.S. 755, 769-70 (2018).
[14] Chiles v. Salazar, 607 U.S. ___, slip op. at 13 (2026) (citing Cohen v. California, 403 U.S. 15 (1971); Holder v. Humanitarian Law Project, 561 U.S. 1 (2010)).
[15] Id. at 15.
[16] Id. at 13 (quoting NAACP v. Button, 371 U.S. 415, 429 (1963)).
[17] Id. at 9, 14 (quoting Rosenberger v. Rector and Visitors of Univ. of Va., 515 U.S. 819, 829 (1995)).
[18] Id. at 13.
[19] Transcript of Oral Argument at 23-24, Chiles v. Salazar, No. 24-539 (Oct. 7, 2025) (Alito, J.).
[20] Chiles v. Salazar, 607 U.S. ___, slip op. at 20 (2026).
[21] Id. at 21.
[22] Id. (quoting Illinois ex rel. Madigan v. Telemarketing Assocs., Inc., 538 U.S. 600, 620 (2003)).
[23] Id.
[24] W. Va. State Bd. of Educ. v. Barnette, 319 U.S. 624, 642 (1943); Rosenberger v. Rector and Visitors of Univ. of Va., 515 U.S. 819 (1995); Iancu v. Brunetti, 588 U.S. 388 (2019).
[25] Chiles v. Salazar, 607 U.S. ___, slip op. at 2 (Jackson, J., dissenting) (2026).
[26] Id. at 20.
[27] Id. at 19-20 (quoting R. Post, Informed Consent to Abortion: A First Amendment Analysis of Compelled Physician Speech, 2007 U. Ill. L. Rev. 939, 950).
[28] Planned Parenthood of Se. Pa. v. Casey, 505 U.S. 833, 884 (1992) (plurality opinion), cited in Chiles v. Salazar, 607 U.S. ___, slip op. at 16 (Jackson, J., dissenting) (2026).
[29] Chiles v. Salazar, 607 U.S. ___, slip op. at 20-21 (Jackson, J., dissenting) (2026). The majority made the same observation, noting that Colorado “disputes none of this; neither does the dissent.” See id. at 13 (majority opinion).
[30] Id. at 23 (majority opinion).
[31] Id. (quoting Rosenberger v. Rector and Visitors of Univ. of Va., 515 U.S. 819, 829 (1995)).
[32] 303 Creative LLC v. Elenis, 600 U.S. 570 (2023).
[33] The Trevor Project, 2024 U.S. National Survey on the Mental Health of LGBTQ+ Young People, at 2, 17 (2024), https://www.thetrevorproject.org/survey-2024; see also Brief for American Psychological Association et al. as Amici Curiae, Chiles v. Salazar, No. 24-539 (2025).
[34] Chiles v. Salazar, 607 U.S. ___, slip op. at 22 (2026) (quoting Nat’l Inst. of Fam. and Life Advocs. v. Becerra, 585 U.S. at 772).
[35] United States v. Skrmetti, 605 U.S. 495 (2025).
[36] Chiles v. Salazar, 607 U.S. ___, slip op. at 20-21 (Jackson, J., dissenting) (2026) (citing Skrmetti, 605 U.S. at 516).
[37] Id. at 2 (Kagan, J., concurring) (quoting Iancu v. Brunetti, 588 U.S. 388, 394 (2019)).
[38] Id. at 1.
[39] Reed v. Town of Gilbert, 576 U.S. 155, 165 (2015).
[40] Chiles v. Salazar, 607 U.S. ___, slip op. at 3-4 (Kagan, J., concurring) (2026) (citing Vidal v. Elster, 602 U.S. 286, 330 (2024) (Sotomayor, J., concurring in judgment)).
[41] Day v. Johnson, 255 P.3d 1064, 1068-69 (Colo. 2011); see also Chiles v. Salazar, 607 U.S. ___, slip op. at 21 (2026).
[42] Movement Advancement Project, “Conversion ‘Therapy’ Laws,” last updated March 30, 2026, https://www.lgbtmap.org/equality-maps/conversion\_therapy.
[43] The Trevor Project, 2024 U.S. National Survey on the Mental Health of LGBTQ+ Young People, at 2, 17 (2024).
[44] Chiles v. Salazar, 607 U.S. ___, slip op. at 33-34 (Jackson, J., dissenting) (2026).
[45] Id. at 35.