Viewpoint Discrimination in Talk Therapy: Conversion-Therapy Bans Applied to Pure Speech Require Searching First Amendment Scrutiny
Commentary on Chiles v. Salazar, 607 U. S. ___ (2026) (Decided March 31, 2026)
I. Introduction
Chiles v. Salazar resolves a recurring constitutional question at the intersection of professional regulation and free speech: when a State bans “conversion therapy” for minors and the “therapy” consists solely of conversation, does the State regulate conduct (professional treatment) or speech (and, if so, does it do so in a viewpoint-discriminatory manner)?
Petitioner Kaley Chiles, a licensed Colorado counselor who provides only talk therapy, brought a pre-enforcement, as-applied First Amendment challenge to Colorado’s 2019 “conversion therapy” ban as applied to her counseling of minors. She alleged that the statute allowed counseling that affirms identity exploration and gender transition but prohibited counseling aimed at reducing same-sex attractions, changing sexual behaviors, or “realigning” gender identity with biological sex—thus skewing what she may say based on the viewpoint expressed.
The lower courts (district court and the Tenth Circuit) agreed Chiles had Article III standing but denied a preliminary injunction, treating the law primarily as regulation of professional conduct with only incidental effects on speech and applying only rational-basis review. The Supreme Court granted certiorari to resolve a circuit conflict on how the First Amendment applies to conversion-therapy bans when enforced against talk therapy.
Key issues
- Classification: Is talk therapy “speech as speech” or regulable “professional conduct” with incidental speech effects?
- Discrimination: Does the statute, as applied, regulate content and discriminate by viewpoint?
- Scrutiny: What level of First Amendment scrutiny is required?
- History/tradition: Does a tradition of medical regulation justify diminished First Amendment protection here?
II. Summary of the Opinion
Holding: Colorado’s conversion-therapy ban, as applied to Chiles’s talk therapy, regulates speech based on viewpoint. The lower courts erred by applying insufficiently rigorous First Amendment scrutiny.
The Court emphasized that Chiles provides only speech. Colorado’s statute allows counseling that expresses acceptance and support for identity exploration and transition while forbidding counseling that attempts to change sexual orientation or gender identity (including attempts to reduce same-sex attraction or alter gender expression). That asymmetry, the Court concluded, is classic viewpoint discrimination.
The Court rejected Colorado’s effort (and the Tenth Circuit’s approach) to relabel the regulated speech as “conduct,” “treatment,” or a “therapeutic modality.” It further rejected the idea that licensed professional status creates a First Amendment “free zone” or a special category of “professional speech” entitled to diminished protection, relying heavily on National Institute of Family and Life Advocates v. Becerra.
Finally, the Court declined Colorado’s invitation to broaden historic exceptions by aggregating traditions of licensing, informed consent, and malpractice into a new category of lesser-protected “substandard care” speech.
III. Analysis
A. Precedents Cited (and How They Shaped the Decision)
1. Core content- and viewpoint-discrimination framework
- Reed v. Town of Gilbert — The Court invoked Reed’s rule that content-based regulations are “presumptively unconstitutional” and ordinarily trigger strict scrutiny (narrow tailoring to a compelling interest). Reed supplied the baseline analytical posture: the default is suspicion, not deference, when law turns on what is said.
- Rosenberger v. Rector and Visitors of Univ. of Va. — Rosenberger provided the key escalation: viewpoint discrimination is a particularly “egregious form” of content regulation that government must almost always avoid. The Court treated Colorado’s statute as doing precisely what Rosenberger forbids—permitting affirmative counseling but prohibiting change-directed counseling on the same subject.
- Iancu v. Brunetti, Good News Club v. Milford Central School, West Virginia Bd. of Ed. v. Barnette — These cases reinforced the prohibition on governmental orthodoxy in ideas and the bedrock rule against viewpoint discrimination.
2. Rejecting “speech is conduct” relabeling
- NAACP v. Button — Button supplied the Court’s anti-label principle: constitutional rights cannot be curtailed by “mere labels.” This was the doctrinal tool used to dismantle Colorado’s “therapy is conduct” framing as applied to a talk-only practice.
- Cohen v. California — Cohen illustrated that even when a statute is nominally aimed at “conduct” (disturbing the peace), it triggers robust First Amendment scrutiny when applied to pure expression. The Court treated Chiles’s case as Cohen-like: the only “conduct” is speech.
- Holder v. Humanitarian Law Project — Holder confirmed that a law with many conduct applications still triggers strict scrutiny when applied to speech (training/advice). The Court used Holder to rebut the Tenth Circuit’s method of looking to the statute’s “main thrust” rather than the as-applied speech restriction.
3. “Professional speech” and the limits of diminished protection
- National Institute of Family and Life Advocates v. Becerra — NIFLA was the central modern precedent. The Court leaned on NIFLA’s rejection of a broad “professional speech” category with diminished constitutional protection. It also adopted NIFLA’s two recognized contexts where heightened scrutiny may be relaxed: (i) compelled disclosure of factual, noncontroversial information in commercial speech (via Zauderer v. Office of Disciplinary Counsel of Supreme Court of Ohio), and (ii) regulation of conduct that only incidentally burdens speech. The Court held Colorado’s law fit neither.
- United States v. Alvarez — Alvarez contributed the “historic and traditional categories” framing and the insistence that new categories of lesser-protected speech are disfavored absent a strong historical pedigree.
- R. A. V. v. St. Paul — R.A.V. supported the proposition that even within certain unprotected categories (like fighting words), viewpoint discrimination remains deeply suspect.
4. The “speech incident to conduct” doctrine (and its boundaries)
- United States v. Williams and Giboney v. Empire Storage & Ice Co. — These cases illustrated one pathway where speech may be regulable without strict scrutiny when integrally related to unlawful conduct (e.g., speech facilitating crime/contraband). The Court found no analogous unlawful conduct here: clients may lawfully pursue behavioral changes; the State’s target is the viewpoint expressed.
- Sorrell v. IMS Health Inc. — Sorrell supplied the contrasting example: viewpoint-neutral regulation of conduct (like an outdoor-fire ban) can incidentally burden expressive acts without triggering strict scrutiny. The Court concluded Colorado’s law is not message-neutral; it “trains directly on” content and viewpoint.
5. Standing and pre-enforcement challenges
- Susan B. Anthony List v. Driehaus — Although not central to the merits, Driehaus anchored the standing analysis: a credible threat of enforcement plus chill of speech can establish an injury for pre-enforcement review. The Court endorsed the lower courts’ standing findings.
- People v. Rodriguez-Morelos — Cited to reject Colorado’s late-breaking narrowing construction as surplusage-creating, reinforcing that standing cannot be defeated by opportunistic reinterpretation that contradicts earlier litigation positions.
6. Colorado’s “history and tradition” effort and why it failed
- United States v. Stevens — Stevens provided the Court’s warning against creating a “First Amendment Free Zone” and against broad, generality-driven categories of unprotected speech. The Court characterized Colorado’s proposed “substandard care” speech category as Stevens-like overbreadth in conceptual form.
- Illinois ex rel. Madigan v. Telemarketing Associates, Inc. — Used to contrast malpractice’s “exacting proof requirements” (injury, causation) that provide “breathing room,” with Colorado’s ex ante licensing discipline that penalizes speech alone without individualized harm proof.
- Day v. Johnson and Schneider v. Revici — Malpractice cases/authorities used to show malpractice’s structure differs from viewpoint-based censorship enforced by licensing sanctions, and that consent may sometimes permit deviation from standard care (unavailable under Colorado’s ban).
- Buck v. Bell — Invoked as a cautionary example about reflexive deference to prevailing professional orthodoxies.
7. Circuit conflict context
- Tingley v. Ferguson and Catholic Charities of Jackson, Lenawee and Hillsdale Ctys. v. Whitmer — Cited in connection with standing and the broader lower-court treatment of similar statutes, illustrating the fractured approaches.
- Otto v. Boca Raton — Used (by the Court) to underline the danger of treating speech-heavy practices as regulable conduct by definition.
- King v. Governor of New Jersey (and its partial abrogation by National Institute of Family and Life Advocates v. Becerra) — Provided the doctrinal backstory: some circuits had used intermediate scrutiny approaches premised on “professional speech” concepts that NIFLA destabilized.
8. The concurrence and dissent as interpretive foils
- Vidal v. Elster and Davenport v. Washington Ed. Assn. — In Justice Kagan’s concurrence, these cases support a more nuanced posture toward some content-based but viewpoint-neutral regimes where censorship risk is attenuated. The concurrence signals a narrower ruling: viewpoint discrimination is the “easy case,” while viewpoint-neutral healthcare speech regulation may raise harder questions.
- Lambert v. Yellowley and Lowe v. SEC — In Justice Jackson’s dissent, these cases support robust state police power to regulate professions and the proposition that regulation of professional practice does not evaporate because speech is involved.
- Planned Parenthood of Southeastern Pa. v. Casey — The dissent’s central analogue: informed-consent speech compelled “as part of the practice of medicine” can be regulated; the dissent argued the majority misread NIFLA and Casey and destabilized healthcare regulation.
B. Legal Reasoning
1. The Court’s threshold move: focus on the as-applied reality
The opinion insists that First Amendment analysis begins with what the law does in the case at hand. Even if Colorado’s statute reaches aversive physical interventions in other applications, the Court treated Chiles’s practice as pure talk therapy. That factual premise collapses the “conduct” framing: “All Ms. Chiles does is speak—and, as far as she is concerned, speech is all Colorado seeks to regulate.”
2. Content discrimination becomes viewpoint discrimination
Colorado’s statute is not merely topic-based. It permits affirming and supportive counseling for identity exploration and gender transition, while forbidding counseling that attempts to change identity or reduce same-sex attraction. On the Court’s reading, this is not a neutral regulation of “dangerous techniques”; it is a state preference about which therapeutic answers may be voiced in the counseling room. That preference is “textbook” viewpoint discrimination.
3. “Professional speech” is not a constitutional carveout
The Court treated NIFLA as foreclosing diminished protection for speech simply because the speaker is licensed. The opinion’s historical references (governments manipulating “doctor-patient discourse” to increase state power or suppress minorities) are used to justify why professional settings are not exempt from distrust of viewpoint-based censorship.
4. The “incident to conduct” doctrine is narrowed to its classic forms
The Court re-articulated two recognized situations where speech restrictions may avoid strict scrutiny as “incident to conduct”:
- Integral-to-unlawful-conduct: speech closely tied to separately unlawful activity (e.g., contraband solicitation/transactions).
- Content-independent regulation of conduct: restrictions that apply without regard to communicative message (e.g., fire bans incidentally restricting flag burning).
Colorado’s law, as applied, did neither: it was not tethered to independently unlawful conduct, and it was not message-neutral. The Court therefore rejected rational-basis (and, implicitly, intermediate-scrutiny) approaches used by some lower courts to sustain similar bans.
5. No new “substandard care” exception through history aggregation
Colorado’s fallback theory asked the Court to recognize a tradition permitting broad censorship of speech deemed “substandard care,” drawing on licensing, informed consent, and malpractice. The Court refused, treating this as the kind of generalized category-building rejected in Alvarez and Stevens. Examined separately, each tradition failed to justify the statute’s viewpoint-based suppression:
- Licensing: too recent as applied to counseling (first counselor-licensure bill in 1976) and historically about qualifications, not mandated viewpoints.
- Informed consent: typically factual/noncontroversial disclosure tied to physical procedures; not viewpoint silencing.
- Malpractice: requires proof of injury/causation and allows more breathing room than licensing discipline triggered by speech alone.
C. Impact
1. Immediate doctrinal consequence: strict scrutiny (or something comparably rigorous) becomes unavoidable for viewpoint-skewed talk-therapy bans
The most direct effect is methodological: courts may no longer uphold the as-applied enforcement of conversion-therapy bans against talk therapists by classifying the speech as “professional conduct” and applying rational-basis review. When the statute is structured to permit affirmation but prohibit change-directed counseling, the Court’s reasoning frames the dispute as viewpoint discrimination—placing the regulation at the apex of First Amendment suspicion.
2. Statutory drafting pressure on States
States wishing to regulate conversion therapy for minors will face incentives to:
- avoid explicit asymmetries that facially privilege one “side” of an ideologically contested question in talk therapy;
- target non-speech conduct (coercion, fraud, threats, or aversive practices) with clearer traditional hooks;
- rely on generally applicable professional-discipline standards that are not viewpoint-selective on their face (while still confronting Chiles’s skepticism of “labeling”).
Justice Kagan’s concurrence underscores that viewpoint-neutral content regulation in clinical settings may present “a different and more difficult question,” leaving open continued litigation over what “viewpoint neutrality” can mean in healthcare regulation.
3. Ripple effects beyond conversion therapy
The reasoning is not limited to LGBTQ-related counseling. Any licensing rule that allows one therapeutic viewpoint but prohibits the opposite viewpoint—where the regulated activity is purely conversational—risks being treated as viewpoint discrimination requiring the most demanding First Amendment review.
4. Litigation posture: as-applied challenges and the “talk-therapy” fact pattern
The Court repeatedly emphasized the narrowness of the question (as-applied to talk therapy; not contesting bans on aversive physical interventions). That framing will likely channel future plaintiffs into:
- as-applied challenges focused on “speech-only” modalities; and
- pre-enforcement challenges grounded in “credible threat” and chill of speech (reinforced by Driehaus).
IV. Complex Concepts Simplified
1. Content-based vs. viewpoint-based restrictions
- Content-based: the government regulates what you talk about (the subject matter).
- Viewpoint-based: the government regulates which side of the subject you may express (permitting one perspective and forbidding the opposite).
The Court treated Colorado’s scheme as viewpoint-based because it permitted affirming counseling while forbidding change-directed counseling on the same topics.
2. “Strict scrutiny”
Strict scrutiny is the Court’s most demanding review standard for speech restrictions: the government must prove a compelling interest and that the law is narrowly tailored (no broader than necessary, and no less-restrictive alternative would achieve the goal).
3. “Speech incident to conduct”
Sometimes a law regulates conduct and speech is affected only as a byproduct (e.g., rules about how to carry out a physical procedure, or a neutral conduct rule that incidentally limits expressive acts). The Court held Colorado’s law, as applied here, was not incidental: it directly targeted what may be said and favored some viewpoints over others.
4. Pre-enforcement standing and “credible threat”
You need not wait to be punished to sue. If you plausibly intend to engage in arguably proscribed speech and face a credible threat of enforcement, the chill itself can be an injury (as recognized in Driehaus and applied here).
V. Conclusion
Chiles v. Salazar establishes (at least as applied to pure talk therapy) a sharp constitutional limit on States’ ability to enforce conversion-therapy bans that select between therapeutic viewpoints. The Court rejected the “professional conduct” label as insufficient to dilute First Amendment scrutiny where the regulated act is speech alone, reaffirmed that there is no broad “professional speech” exception, and refused to create a new “substandard care” category of lesser-protected expression by stitching together disparate historical traditions.
The decision’s practical significance is twofold: it alters the litigation terrain for conversion-therapy restrictions nationwide, and it signals heightened judicial skepticism toward any professional-licensing regime that attempts to enforce an orthodoxy of ideas through viewpoint-selective controls on what licensed professionals may say in private counseling conversations.
Comments