The federal layer is loud. The state layer is faster.
When a telehealth brand thinks about regulatory risk on advertising, the instinct is to think federally. FDA on drug claims. FTC on deceptive practices. Maybe DEA on controlled substances.
That mental model is increasingly out of date.
In 2026, the most active and the most operationally disruptive enforcement against DTC telehealth advertising is coming from state attorneys general. Not in coordinated multi-state actions. Not in slow federal rulemaking. In short, sharp, public moves that put a brand or a platform on the defensive within a news cycle.
In March 2026, Connecticut Attorney General William Tong publicly pressured Meta to take action on what he described as misleading AI-generated weight loss advertisements running on Instagram and Facebook. The letter named specific patterns: deepfake-style celebrity endorsements, AI-rewritten testimonials, and synthetic before-and-after imagery on weight loss compounded GLP-1 ads.
Similar inquiries have followed from California and New York, focused on AI-generated medical content and unsubstantiated health claims. Texas has moved on telehealth advertising from a separate angle, focused on identity verification and prescriber disclosure.
This is not coordinated. It is parallel. And that is the part that matters operationally.
A state AG can demand documents, subpoena ad libraries, name brands in press releases, and refer to other AGs. They do not need a federal trigger. They do not wait for FTC. They move on consumer complaints, news coverage, and political timing.
For DTC telehealth, this is the new center of advertising risk.
This post is how marketing, legal, and operations should think about state AG action specifically, and how to audit your current creative and process before the letter shows up.
Why state AGs are leading on telehealth ads
A few structural reasons.
| Driver | Why it matters |
|---|---|
| State consumer-protection statutes are broad | "Unfair or deceptive acts" laws give state AGs wide authority over advertising, often broader than the FTC Act |
| Statutory damages and per-violation penalties | Many state UDAP statutes carry per-violation penalties that scale with the number of impressions or recipients |
| Political visibility | AGs are elected. High-profile consumer protection cases are political wins |
| Federal slowness | FTC takes years to bring a major case. AGs can move in weeks |
| AI imagery is concrete and provable | A deepfake celebrity endorsement is easy for a regulator to point at |
| Telehealth complaints rose sharply in 2025-2026 | The complaint volume gives AGs an empirical basis to act |
| Multi-state coordination is easy | One AG action sets a template other states adopt |
The last point is the most operationally important. A finding by Connecticut against a brand effectively becomes a template letter that the next state AG can drop on the same brand with light modification.
For brands that operate in 30 to 50 states, the risk is not one inquiry. It is a cascade.
What the recent actions actually targeted
The substance of the state AG concerns, distilled.
AI-generated celebrity endorsements
Deepfake-style ads using celebrity likeness, voice, or fabricated quotes promoting weight loss or other medical products. This is the cleanest type of case for an AG: it combines false advertising, right of publicity violations, and consumer deception in one creative.
Even when the brand is not the direct creator (an affiliate ran the ad), regulators are increasingly comfortable holding the brand accountable for what its affiliates and aggregators run.
AI-rewritten or fabricated testimonials
Patient quotes that were generated, paraphrased, or composited by AI, presented as real testimonials. Also patient quotes that are real but have been altered for tone or specificity.
The FTC's amended Endorsements Guide already covers this at the federal level. State AGs apply parallel state UDAP statutes, often with steeper per-violation penalties.
Synthetic before-and-after photos
Photos that were generated, retouched beyond reason, or composited from multiple sources, presented as actual patient outcomes. This is particularly heavily targeted in the GLP-1 weight loss category.
Unsubstantiated health claims
Claims about weight loss amounts, cardiovascular benefit, anti-aging effects, or chronic disease reversal that lack adequate substantiation. The state law standard for substantiation in health claims is generally high, and what passes a legal claims review at one brand may not pass an AG inquiry.
Off-label or pre-approval marketing
Ads that promote uses of a drug not yet FDA approved, or that mention compounded versions in ways that imply equivalence with branded products. The compounded GLP-1 crackdown made this a particularly visible category.
Inadequate disclosure of prescriber relationships and affiliate compensation
Ads run by affiliates or creators without clear disclosure of the brand relationship. AGs have shown willingness to act on this independent of FTC's affiliate rules.
For the broader compliance angle on affiliate creator programs, see Affiliate and Creator Programs for DTC Telehealth: How to Grow Without FTC, Fake Review, or Medical-Claims Risk.
The targeting pattern
State AGs are not selecting brands at random. The pattern that triggers inquiry is identifiable.
| Risk factor | Why it triggers attention |
|---|---|
| High consumer complaint volume in a state | Direct empirical basis for action |
| Prominent ad spend in the state | High visibility, high impressions |
| Negative news coverage or social-media virality | Political pressure to respond |
| Industry-wide enforcement moment (compounded GLP-1, AI ads) | Brands in the category become targets |
| Specific creative pattern flagged by another state | Template letter potential |
| Operating in the state without state-specific provider licensure | Adds a clear legal predicate |
| Refund and chargeback patterns visible to regulators | Often the first place an inquiry starts |
The complaint-volume pattern is the one most brands underestimate. State AG offices field consumer complaints. When complaints cluster around a brand, an analyst opens a file. That file may sit for months before anything happens publicly. Then it lands as a CID, a subpoena, or a published letter.
A brand that does not know which states generate its complaints is operating blind to the risk.
The creative audit your team should run this month
The brands that survive AG scrutiny share a pattern: they ran the audit before the AG did.
Here is a working framework. Run it across your active ad accounts, affiliate creatives, and landing pages.
AI-generated assets
For every image, video, and audio asset in active campaigns:
- Was any portion generated, altered, or composited by AI?
- If yes, does the asset still accurately represent the product, patient, or outcome?
- Does any AI-generated likeness resemble a real public figure or patient?
- Is the AI generation disclosed where required by state law (California's AB 730 and similar)?
The simplest decision rule: if an asset depicts an outcome, a patient, or a person, it should not be AI-generated unless the AI generation is clearly disclosed and the depicted facts are independently verified.
Testimonials and endorsements
For every quoted patient, creator, or expert:
- Is the person real, identifiable, and contactable?
- Did they actually give the quoted statement?
- Was the statement edited beyond minor formatting?
- Is the relationship to the brand disclosed (compensation, free product, employee)?
- Does the testimonial imply a typical result, and is that typical result substantiated?
The FTC Endorsements Guide is the federal floor. State AGs apply state UDAP statutes that frequently impose stricter standards on what constitutes a "typical" experience claim.
Before-and-after photos
For every weight loss, hair loss, sexual health, dermatology, or other outcome image:
- Is the patient real?
- Was the photo taken under consistent conditions (lighting, distance, pose, clothing)?
- Has the photo been retouched beyond color correction?
- Is the displayed timeframe accurate?
- Is the use authorized in writing for the specific channels running it?
A program that cannot produce written authorization for every before-and-after asset has an unmanaged risk surface.
Claim substantiation
For every health, outcome, or comparative claim in copy:
- Is there substantiation on file?
- Does the substantiation match the specificity of the claim?
- Was the substantiation reviewed in the last 12 months?
- For compounded drugs, are claims consistent with what is permissible for compounded products specifically?
The standard pattern that fails: a claim referencing a published study that does not actually support the specific framing in the ad copy.
Disclosure and transparency
Across the creative pool:
- Are required disclosures legible and proximate to the claim?
- Are affiliate and creator relationships disclosed?
- Are prescriber relationships disclosed where applicable?
- Are the conditions of refund, subscription, and renewal disclosed in a way a regulator would consider conspicuous?
Process gaps that get brands in trouble
Most enforcement actions are not about a single bad ad. They are about a process that produced many bad ads, or that could not catch a bad one.
The process patterns that draw AG attention:
No central creative review
Affiliates, agencies, and in-house teams ship ads without consolidated review. The brand cannot say with confidence what is running.
No ad library audit cadence
Brands run ads that pre-date current policy and never re-review. Old creative with outdated claims keeps spending.
No documented creative approval
Approvals happen in chat, email, or unrecorded calls. When asked who approved a specific asset, no one can answer.
Affiliate creative not in scope
Affiliate and creator content is treated as outside the brand's responsibility. State AGs and the FTC disagree.
Complaints handled but not analyzed
Customer support resolves complaints, but no one is reading the patterns. A state-by-state complaint heatmap is the single most useful early-warning signal for AG attention.
Refund and chargeback patterns invisible
Spikes in refunds or chargebacks in a state precede AG inquiries more often than they coincide with them.
For the broader connection between refund patterns and trust, see Reducing Refunds and Chargebacks in Subscription Telehealth.
What to do if a state AG inquiry lands
The first hours and days matter.
Do not respond off the cuff
A state AG inquiry letter or CID is not the place to draft a casual reply. Engage counsel immediately. The response sets the tone and scope of the inquiry.
Preserve everything
The moment an inquiry lands, preserve all advertising assets, approval records, complaint logs, affiliate communications, and refund records in the relevant time window. Document a litigation hold internally.
Map the scope honestly
Understand what the AG actually asked for. Many inquiries start broad and narrow. Responses that volunteer information beyond the request expand the scope.
Audit your own creative first
Whatever the AG asked about, assume they will find more if they look harder. The brands that respond well already know what would be found and have a clean story about why and how it was happening.
Run a unified response across states
If one state acts, others will follow. A coordinated response posture, with consistent positions, prevents being whipsawed by parallel inquiries.
Decide on remediation before being asked
Voluntary remediation, documented before the AG requires it, can be the difference between a public settlement and a quiet resolution. Pulling specific creative, changing affiliate terms, or tightening review processes are common pre-emptive remediations.
Operational changes to make before an inquiry
Some changes are worth making now, AG inquiry or not.
Central creative inventory
A single source of truth for every ad asset running anywhere on any channel, including affiliate creative. With approval status, version history, and expiration.
Creative approval workflow with audit trail
Every asset goes through a documented review (legal, regulatory, clinical where applicable). Approvals are recorded against the asset and exportable.
Affiliate creative pre-approval
Affiliates submit creative before running it. The brand maintains the right to require changes. Compensation can be conditioned on compliance.
State-by-state complaint dashboard
Customer support tags complaints by state. A weekly review highlights any state where complaints spike.
Quarterly creative re-review
Every active asset is re-reviewed every quarter, with claims substantiation re-validated.
AI-asset register
Any AI-generated or AI-modified asset is logged with provenance, generation tools, and disclosure status.
Right-of-publicity diligence
For any person depicted, real or AI-generated, the brand has a documented authorization or a clear basis for not needing one.
For the broader marketing operations layer, see Marketing Your GLP-1 Program in 2026 and Running GLP-1 Ads in 2026.
How this fits with federal and platform layers
The state AG layer does not replace federal or platform enforcement. It adds to it.
| Layer | What it catches | Typical timeline |
|---|---|---|
| Platform policy (Meta, Google, TikTok) | Disallowed creative, restricted categories | Hours to days |
| State AG | Deceptive practices under state UDAP, false advertising | Weeks to months |
| FTC | Federal deceptive practices, endorsement violations | Months to years |
| FDA | Misbranding, off-label promotion, unapproved claims | Months to years |
| Private class action | Consumer or competitor lawsuit | Months to years |
A brand that has clean creative passes all five. A brand that fails platform policy is the first warning sign that other layers are coming.
For ad-platform-policy specifics, see Running GLP-1 Ads in 2026 and Instagram Ads for GLP-1 Programs.
Implementation checklist
Use this when scoping a creative compliance program.
Inventory
- Central register of every active ad asset across channels
- Affiliate and creator creative included in the register
- Version history and expiration dates per asset
Review
- Documented approval workflow with clinical, legal, regulatory review
- Affiliate creative pre-approval enforced contractually
- Quarterly re-review of every active asset
AI specifically
- AI-asset register with provenance and disclosure status
- Decision rule that AI-generated outcome imagery is not used
- Disclosure language consistent with state requirements
- Right-of-publicity check for any AI likeness
Testimonials and endorsements
- Real, identifiable, contactable people only
- Written authorization for every quoted patient or creator
- Relationship disclosure on every endorsement
- Typical-result substantiation where required
Before-and-after
- Real patient, written authorization on file
- Conditions documented (lighting, distance, timeframe)
- No retouching beyond color correction
Complaints and risk monitoring
- State-tagged complaint dashboard reviewed weekly
- Refund and chargeback patterns reviewed by state
- Spike alerts in place
Response readiness
- AG inquiry response playbook documented
- Litigation-hold process defined
- Counsel engagement on retainer or pre-identified
- Coordinated cross-state response posture
Final takeaways
State AG action is the layer of telehealth advertising enforcement that moves fastest, hits hardest, and is most likely to cascade.
What to remember:
- The federal layer is not the threat anymore; state AGs are
- AI-generated outcome imagery, deepfake endorsements, and fabricated testimonials are the cleanest categories to target
- Targeting is patterned: complaint volume, ad spend, news visibility, and industry-wide moments
- One AG action becomes a template for the next state
- The first defense is a documented creative review process, not a brilliant lawyer at the back end
- The second defense is a complaint and refund dashboard that lets you see risk by state before regulators do
- The third defense is voluntary remediation before being asked
- Brands that survive scrutiny ran the audit before the AG did
The brands that will spend the next two years confidently advertising in this category are the ones that take a quiet week now to clean up their creative, document their process, and turn complaint data into an early-warning system.
The brands that wait until the letter arrives will spend that time in a much harder conversation.
Sources for the regulatory and case references:
- Connecticut Attorney General William Tong Press Release on Meta AI Weight Loss Ads
- FTC Endorsement Guides Final Rule
- FTC Final Order Against NextMed - FTC
- California AB 730 Deepfakes and Synthetic Media
- State Consumer Protection Statutes Overview - National Consumer Law Center
- FDA Warned 30+ Telehealth Companies in 2026 - Health Data Consortium