New York's AI Video Disclosure Law: What the Synthetic Performer Mandate Means for Advertisers, Brands, and Content Creators in 2026
New York's groundbreaking AI disclosure law (S8420A) requires advertisers to label AI-generated synthetic performers starting June 2026. Understand compliance requirements, penalties, exemptions, and how to implement disclosure across Meta, Google, TikTok, and CTV.
On December 11, 2025, New York Governor Kathy Hochul signed S.8420-A/A.8887-B into law—the first state legislation in the nation requiring advertisers to explicitly disclose when AI-generated "synthetic performers" appear in commercial advertisements. Effective June 9, 2026, the law mandates conspicuous disclosure across all paid advertising channels including Meta, Google, YouTube, TikTok, CTV, and display networks.
This legislation represents a watershed moment in AI regulation. While federal frameworks debate broad AI governance, New York has taken decisive action with a narrowly targeted, enforceable mandate that affects every advertiser, brand, agency, and content creator doing business in the state. Violations carry civil penalties of $1,000 for first offenses and $5,000 for subsequent violations—and with New York's aggressive consumer protection enforcement history, compliance is not optional.
Alongside the disclosure law, New York simultaneously enacted S.8391, expanding post-mortem publicity rights to cover digital replicas of deceased performers. Together, these laws establish New York as the most stringent jurisdiction for AI-generated human imagery in advertising, setting precedents likely to cascade across California, Illinois, Texas, and eventually federal legislation.
This comprehensive guide explores the legal requirements, definitions, penalties, exemptions, industry impact, compliance strategies, and technical solutions—including face blurring tools like bgblur.com—to help advertisers navigate this new regulatory landscape.
Part I: The Law – S.8420-A Explained
What is S.8420-A?
Full Title: An act to amend the general business law, in relation to requiring disclosure of the use of artificial intelligence in advertisements
Signed: December 11, 2025 by Governor Kathy Hochul
Effective Date: June 9, 2026 (180 days after signing)
Amends: New York General Business Law § 396-b
Legislative Intent
Governor Hochul's signing statement emphasized:
"As artificial intelligence continues to reshape the advertising industry, New Yorkers deserve transparency about when they're viewing AI-generated content versus real human performers. This law protects consumers from deceptive marketing practices while preserving creative freedom for legitimate artistic expression."
The law responds to three converging trends:
Proliferation of generative AI tools - Midjourney, Stable Diffusion, Runway, Pika, and others make synthetic human imagery trivially easy
Consumer confusion - Studies show 67% of consumers cannot reliably distinguish AI-generated faces from real photos
Deepfake abuse in advertising - Cases of unauthorized AI-generated celebrity endorsements, fake testimonials, and deceptive product claims
"Any person engaged in the business of dealing in any property or service who, for any commercial purpose, produces or creates an advertisement"
Practical scope:
✅ Advertisers - Brands, businesses, e-commerce sellers
✅ Advertising agencies - Creative agencies producing ad content
✅ Marketing platforms - Companies creating ad campaigns for clients
✅ Political campaigns - Candidates and PACs running paid ads
✅ Influencers/creators - Monetized sponsored content
❌ Media outlets/platforms - Newspapers, magazines, TV networks, streaming services, billboards, and social platforms that publish or disseminate ads are not liable (similar to Section 230 immunity)
What is a "Synthetic Performer"?
The law defines a synthetic performer as:
"A digital asset that is created, reproduced, or modified by computer, using generative artificial intelligence or a software algorithm, that is intended to give the impression that the asset is in an audio, audiovisual, and/or visual performance of a human performer when it is not recognizable as any identifiable natural performer."
Key elements:
Digitally created or modified - Using generative AI or software algorithms
Impression of human performance - Looks/sounds like a real person
Not recognizable as any identifiable person - Not a specific celebrity or individual
Examples that qualify as synthetic performers:
AI-generated faces created with Midjourney/Stable Diffusion that appear in product ads
Synthetic voiceovers created with ElevenLabs/Descript that narrate commercials
Virtual influencers like Lil Miquela or Aitana Lopez promoting products
Composite images blending multiple real faces into a non-identifiable person
Deepfake performances where an AI-generated person demonstrates a product
Examples that do NOT qualify:
Real human models photographed for ads (even if Photoshopped)
Identifiable celebrities - Because they are "recognizable as an identifiable natural person," separate laws (like S.8391 on digital replicas) apply
Cartoon/animated characters - Not intended to give impression of real human
Product-only images - No human performers at all
Background/wide shots - Where humans are incidental, not the performer
What Does "Conspicuous Disclosure" Mean?
The law requires disclosure to be "conspicuous" but does not specify:
Exact wording/language
Placement (overlay, caption, disclaimer)
Size/font requirements
Format differences by medium (video vs static vs audio)
Legal standard for "conspicuous":
New York courts typically interpret "conspicuous" to mean:
Visible/audible - Not hidden in fine print or buried in disclosures
Proximate - Near the synthetic performer, not at end of 30-second ad
Clear - Understandable language, not legalese
Proportional - Appropriately sized for the medium
Practical implementations emerging:
For Video Ads (Meta, YouTube, TikTok, CTV):
Text overlay: "This ad features AI-generated performers"
Corner badge: "Synthetic Performer" icon
Voiceover disclaimer: "The people in this ad are AI-generated"
For Static Ads (Display, Instagram, Facebook):
Corner watermark: "AI Generated"
Caption text: "Features synthetic performers"
Hashtag: #AIGenerated #SyntheticPerformer
For Audio Ads (Spotify, Podcasts):
Verbal disclaimer: "This advertisement features a synthetic voice"
Disclosure is required only if the advertiser has "actual knowledge" that a synthetic performer is used.
What constitutes "actual knowledge"?
✅ You generated the image with Midjourney
✅ Your agency told you it's AI-generated
✅ The vendor invoice lists "AI model generation"
✅ You explicitly requested synthetic performers
❓ You suspect but didn't verify
❌ You purchased stock photos without asking
❌ A third-party vendor handled creative without disclosure
Practical implication:
Advertisers may be incentivized to avoid learning whether content is AI-generated. However, New York's consumer protection enforcement typically interprets "actual knowledge" to include willful blindness—deliberately avoiding information you should reasonably know.
Best practice: Require vendors/agencies to disclose AI usage in contracts and deliverables.
Part III: Penalties and Enforcement
Civil Penalties
First violation: $1,000 per advertisement
Subsequent violations: $5,000 per advertisement
What counts as "per advertisement"?
Each unique creative likely counts separately
Same creative across multiple platforms may count as one ad or multiple (unclear pending enforcement)
Example scenario:
An e-commerce brand runs:
5 different AI-generated model photos on Meta
Same 5 images on Google Display
No disclosure on any
Potential penalty:
If counted per unique creative: 5 violations × $1,000 = $5,000
If counted per platform placement: 10 violations × $1,000 = $10,000
If the brand continues running undisclosed ads after the first violation, subsequent penalties jump to $5,000 each.
Enforcement Mechanism
The law is enforced by the New York Attorney General under General Business Law § 396-b, which prohibits deceptive advertising practices.
Enforcement process:
Investigation - AG receives complaints or proactively monitors advertising
Notice - Advertiser receives notice of alleged violation
Cure period - Opportunity to add disclosure and cease violations
Penalty assessment - AG assesses civil penalties
Settlement or litigation - Advertiser may settle or contest in court
Precedent from NY AG enforcement:
New York's AG has aggressively enforced consumer protection laws:
2024: Sued multiple supplement brands for deceptive health claims
2025: Investigated AI-generated fake reviews on e-commerce platforms
Takeaway: NY AG Letitia James has a track record of swift, aggressive enforcement. Compliance is not optional.
Private Right of Action?
No. The law does not create a private right of action—only the AG can enforce.
However, class action lawsuits under existing consumer protection statutes (e.g., New York General Business Law § 349 – deceptive practices) remain possible if consumers can prove they were deceived by undisclosed synthetic performers.
Part IV: Exemptions and Safe Harbors
1. Expressive Works Exemption
Exempt: Advertisements and promotional materials for expressive works (movies, TV shows, streaming content, documentaries, video games) are exempt if the use of synthetic performers in the advertisement is consistent with their use in the underlying work.
Why this exemption exists:
Trailers, posters, and promotional content for films/games often feature CGI, visual effects, and digital performers. Requiring disclosure for every Marvel movie trailer with digital effects would be impractical.
Examples:
✅ Exempt: Movie trailer for a sci-fi film featuring CGI aliens
✅ Exempt: Video game ad showing in-game characters
✅ Exempt: TV show promo with visual effects used in the show
❌ Not exempt: Ad for a movie using AI-generated fake audience testimonials
❌ Not exempt: Video game ad using AI-generated influencer endorsement
Key condition: The synthetic performer use in the ad must match its use in the actual work.
2. Language Translation Exemption
Exempt: Use of AI solely for language translation of a human performer does not trigger disclosure.
Examples:
✅ Exempt: Real actor's dialogue translated to Spanish using AI voice synthesis
✅ Exempt: Lip-sync AI to match dubbed language
❌ Not exempt: AI-generated voice replacing human voice entirely (not translation)
3. Media Outlet/Platform Immunity
Exempt: Media outlets and platforms that publish or disseminate advertisements (newspapers, magazines, TV networks, streaming services, billboards, Meta, Google, TikTok) are not liable for third-party advertiser violations.
Rationale: Similar to Section 230 immunity, platforms are not publishers of third-party content.
Practical implication:
Meta/Google won't reject ads without disclosure (they're not liable)
Advertisers bear full compliance responsibility
Platforms may implement optional disclosure tools (similar to "Paid Partnership" tags)
4. Non-Human Imagery
Exempt: Product-only images or backgrounds without human performers do not require disclosure.
❌ Not exempt: AI-generated human hands holding product
❌ Not exempt: AI-generated person wearing clothing
Part V: The Companion Law – Digital Replicas of Deceased Performers (S.8391)
Overview
Simultaneously with S.8420-A, Governor Hochul signed S.8391—expanding New York's right of publicity law to cover digital replicas of deceased performers.
Effective: December 11, 2025 (immediate)
Who is Protected?
"Deceased personalities" - Natural persons:
Domiciled in New York at time of death
Whose name, voice, signature, photograph, or likeness had commercial value at death or gained such value posthumously
"Computer-generated, electronic representations that are readily identifiable as the voice or visual likeness of an individual in sound recordings, images, audiovisual works, or transmissions where the actual individual did not perform or where the performance has been materially altered."
Examples:
AI-generated deepfake of James Dean promoting a product
Synthetic voice of Aretha Franklin in a commercial
CGI recreation of Robin Williams in an ad
Consent Requirements
Unauthorized use of a deceased personality's:
Name
Voice
Signature
Photograph
Likeness
Digital replica
for commercial purposes requires consent from specified rights holders:
Deceased's estate/trust
Surviving spouse
Children
Parents
Siblings (in order of priority)
Consent duration: Rights last for 40 years after death.
Penalties
Damages:
Greater of $2,000 or actual compensatory damages
Attributable profits from unauthorized use
Punitive damages (if willful)
Knowledge requirement: Liable only if user knew the use was unauthorized.
Strategic Implications
This law directly targets the "Deepfake Tupac endorses crypto" problem—unauthorized AI resurrections of deceased celebrities for commercial gain.
Impact on advertisers:
Cannot use AI-generated celebrity likenesses without estate approval
Estate consent costs will rise (scarcity of available celebrities)
Increased legal risk for vintage celebrity campaigns
Example case:
An energy drink brand uses AI to generate a deepfake Michael Jackson performance in a commercial. Jackson's estate files suit under S.8391:
Statutory damages: $2,000 minimum
Actual damages: Market rate for MJ endorsement (~$5-10M)
Attributable profits: Campaign drove $50M in sales
Total exposure: $10M+ damages
Part VI: Industry Impact Analysis
1. Digital Advertising (Meta, Google, TikTok, Snap)
Current state:
AI-generated model images are ubiquitous in e-commerce advertising:
Fashion brands use Midjourney models to avoid photoshoot costs
Supplement brands generate diverse body types for A/B testing
Local businesses create fake customer testimonials
Impact of S.8420-A:
✅ Compliant approach:
Add "AI-generated model" disclosure to creative
Update ad templates with disclosure overlays
Train media buyers on compliance
❌ Risky approach:
Continue using synthetic performers without disclosure
Hope for lack of enforcement (bad bet in NY)
Operational changes:
Current Workflow
Post-June 2026 Workflow
Generate AI model image
Generate AI model image
Upload to Meta Ads
Add disclosure overlay
Launch campaign
Document AI usage
—
Launch campaign
Cost impact: Minimal—adding disclosure text/overlay takes <5 minutes per creative.
Conversion impact: Unknown—will "AI-generated" disclosure reduce click-through or conversion rates?
While S.8420-A applies to commercial advertising, New York's amended election law separately mandates disclosure of "materially deceptive media" in political communications.
Definition: Political communication produced by or including materially deceptive media must disclose if the distributor has actual knowledge.
Examples:
Deepfake video of candidate saying something they never said
AI-generated crowd at rally
Synthetic voice in attack ad
Disclosure requirement: Clearer and more prominent than commercial ad disclosure.
Penalties: Civil penalties plus potential criminal charges for election fraud.
Impact on 2026 midterms:
New York political advertisers must:
Disclose any AI-generated imagery in campaign ads
Verify vendor compliance
Document disclosure in FEC filings
4. E-commerce & Direct-to-Consumer Brands
Current state:
DTC brands heavily use AI-generated models for:
Product photos (clothing, accessories)
Lifestyle imagery
User-generated content (fake reviews)
Impact:
Scenario 1: Fashion brand uses AI models
Before:
Generate 50 AI model images for new collection
Upload to Shopify product pages
Run Meta ads
After:
Generate 50 AI model images
Add "Model is AI-generated" watermark to each image
Must disclose "Testimonial features synthetic performer"
(Or switch to real customer testimonials)
Strategic shift:
Many DTC brands are returning to human models to avoid disclosure friction. Cost savings from AI generation are offset by:
Reduced trust from disclosure
Legal compliance overhead
Potential conversion rate decrease
5. Film, TV, and Gaming Industries
Impact:Minimal due to expressive works exemption.
Movie trailers, TV promos, and game ads featuring CGI/VFX are exempt as long as synthetic performer use matches the underlying work.
However: Promotional content using AI-generated audience testimonials, fake critic quotes, or synthetic influencer endorsements is not exempt and requires disclosure.
Part VII: Compliance Strategies
Step 1: Audit Existing Advertising
Action items:
Inventory all active campaigns across Meta, Google, TikTok, display, CTV, etc.
Identify AI-generated content:
Review creative files
Check vendor invoices
Ask designers/agencies
Flag campaigns requiring disclosure
Documentation template:
Campaign ID
Platform
Creative Type
AI-Generated?
Disclosure Added?
FB-2026-01
Meta
Model photo
Yes (Midjourney)
❌ To Do
GG-2026-05
Google
Product only
No
N/A
TT-2026-08
TikTok
Testimonial
Yes (Synthesia)
❌ To Do
Step 2: Implement Disclosure
Option A: Text Overlay
For video ads:
[Corner of frame, first 3 seconds]
"This ad features AI-generated performers"
[or]
"Synthetic performer"
For static ads:
[Top-right corner, white text on semi-transparent background]
"AI Generated"
Option B: Voiceover Disclaimer
For video/audio ads:
[First 3 seconds of audio]
"The following advertisement features AI-generated voices and performances."
Option C: Platform-Specific Disclosure
Some platforms may introduce native disclosure tools (similar to Instagram's "Paid Partnership" tag).
As of June 2026:
Meta: No native tool yet (use text overlay)
Google: No native tool (use text overlay)
TikTok: Testing "AI Content" label (opt-in)
Step 3: Update Contracts and Vendor Agreements
Require vendors/agencies to:
Disclose AI usage in all deliverables
Add compliant disclosure to creative files
Indemnify advertiser for non-compliance
Sample contract clause:
"Vendor warrants that all creative deliverables will include conspicuous disclosure if synthetic performers are used, in compliance with New York General Business Law § 396-b. Vendor agrees to indemnify Client for any penalties, damages, or costs arising from Vendor's failure to comply with applicable AI disclosure laws."
Step 4: Train Internal Teams
Media buyers, designers, and marketing managers must:
Understand S.8420-A requirements
Recognize synthetic performers
Implement disclosure workflow
Document compliance
Training checklist:
✅ What is a synthetic performer?
✅ When is disclosure required?
✅ How to add disclosure (templates)
✅ Exemptions (expressive works, translation)
✅ Penalties for non-compliance
✅ Vendor compliance verification
Step 5: Monitor Regulatory Guidance
The New York Attorney General is expected to publish:
Part VIII: Technical Solutions for Privacy Compliance
Face Blurring for Privacy Protection
While S.8420-A focuses on disclosure of synthetic performers, many advertisers are also concerned about privacy compliance when using real human imagery—especially under laws like:
GDPR (Europe) - Requires consent for identifiable faces
CCPA/CPRA (California) - Regulates biometric data
BIPA (Illinois) - Strict biometric privacy law
Solution: Face blurring tools like bgblur.com enable advertisers to:
✅ Anonymize real people in B-roll footage
✅ Protect employee/customer privacy in case studies
✅ Comply with GDPR without explicit consent
✅ Reduce legal risk from incidental captures
[Workflow]
1. Capture video footage with real people
2. Upload to bgblur.com
3. Automated face detection
4. Apply blur to non-consenting individuals
5. Export final video
6. Use in advertising (no synthetic performers, no disclosure needed)
Strategic question: Should brands embrace or hide AI usage?
Embrace AI (proactive disclosure):
Brands like The North Face and Coca-Cola are experimenting with "AI-native" campaigns that celebrate AI generation:
Ads explicitly branded as "Created with AI"
Behind-the-scenes content showing AI workflow
Educational messaging about AI capabilities
Result: Positioning as innovative, transparent, tech-forward.
Minimize AI (compliance-only disclosure):
Traditional brands add minimal, legally compliant disclosure without drawing attention:
Small text overlay
End-of-ad disclaimer
Hashtags buried in caption
Result: Compliance without changing brand perception.
Future of Advertising
Long-term prediction:
By 2028-2030, AI-generated content will be so ubiquitous that disclosure becomes:
Normalized - Consumers expect most digital ads to include AI
Inverted - "Real human model" becomes a premium signal (like "organic" for food)
Irrelevant - Consumers stop caring about AI vs. real distinction
Analogy:
Today's disclosure debate mirrors the 2000s debate over "Photo retouched" disclaimers. Initially controversial, now consumers assume all fashion imagery is edited. Disclosure faded into irrelevance.
Strategic implication:
Early compliance and transparent communication position brands for long-term trust as AI becomes standard.
Conclusion: Preparing for June 9, 2026
New York's synthetic performer disclosure law represents a turning point in AI regulation—moving from aspirational frameworks to concrete, enforceable mandates with real penalties.
For advertisers, brands, agencies, and content creators doing business in New York, the path forward is clear:
Immediate actions (before June 9, 2026):
✅ Audit all advertising for synthetic performers
✅ Implement disclosure in compliant formats
✅ Update vendor contracts to require AI disclosure
✅ Train internal teams on compliance requirements
✅ Monitor AG guidance for official disclosure templates
Strategic decisions:
Embrace transparency or minimize disclosure?
Continue synthetic performers or return to real humans?
Geo-target compliance or apply nationwide?
Technical solutions:
Face blurring tools like bgblur.com for privacy compliance
Disclosure templates for video, static, and audio ads
Documentation systems to prove compliance
The law's impact extends beyond New York. As California, Illinois, and other states advance similar legislation, multi-state compliance will become the norm. Advertisers who build transparent, scalable AI disclosure workflows today will be well-positioned for the patchwork of state laws emerging tomorrow.
The era of undisclosed AI-generated advertising is ending. The question is not whether to comply, but how to do so in a way that preserves brand trust, creative flexibility, and competitive advantage.
