Meta's AI-Generated Content Policy
After AI image and video generation tools went mainstream, Meta has steadily rolled out and tightened policies on AI content in ads. Here's what advertisers need to know.
The Four Core Principles
1. Disclosure Requirement
- AI-generated images and videos must be flagged at upload
- Ads Manager provides an "AI-generated" checkbox
- Some labels may also be shown to end users
2. Deepfake Blocking
- AI-generated faces of real people (especially public figures) are completely prohibited
- Automatic detection of AI content involving politicians and celebrities
- If caught, the ad is blocked and the account gets a warning
3. Brand Safety Enforcement
- AI-generated content is held to the same Brand Safety standards as everything else
- Inappropriate backgrounds or contexts are filtered automatically
4. Strict on Factual Claims
- AI-generated "proof photos" or "testimonials" are prohibited
- Using AI images as "real customer results" violates policy
Pre-Launch Checklist
Before uploading an AI creative:
- [ ] Check the AI-generated box in Ads Manager
- [ ] Verify whether the creative includes real people (only virtual characters are OK)
- [ ] Review components for potential policy violations
- [ ] Check against both your brand tone and Meta's safety standards
Avoiding policy flags:
- No AI-generated celebrity faces (including composites)
- No false "trusted institution" settings like hospitals or labs
- No fake proof images along the lines of "use this product and you'll get X"
Concrete Examples
OK:
- AI-generated virtual character lifestyle images
- Product shots composited with AI backgrounds
- Abstract AI artwork
- AI-generated stock photos (no real photography)
NOT OK:
- AI composites of famous actors' faces
- AI-generated people claiming to be "doctors" or "experts" (when they're AI)
- AI-generated before-and-after photos (claimed as real users)
- Distorted depictions of competitor products via AI
Region-Specific Legal Notes
Extra caution for advertisers in regulated markets:
- Personal data laws: Even AI-generated, if the image resembles a real person, likeness rights risk applies
- Fair advertising laws: Failure to disclose AI use can be deemed unfair advertising
- Food and medical: AI-generated "results" images are often classified as exaggerated advertising
Local ad review bodies generally recommend AI disclosure as well.
What Meta Detects Automatically
Detection tech:
- C2PA metadata (Content Credentials standard)
- SynthID (Google) and watermarking
- AI-specific pattern detection (facial asymmetry, finger count, etc.)
- Perceptual hashing (comparison against known AI images)
What happens after detection:
- Automatic label addition
- Ad rejection in serious cases
- Restrictions on repeat-offender accounts
Safe Patterns for AI Creatives
Safe prompt directions:
- "virtual character", "illustration style", "anime character"
- Styles that clearly signal non-realism
- Product- or scene-centric (minimize people)
Prompts to avoid:
- "realistic photo of doctor"
- "real celebrity [name]"
- "before and after weight loss photo"
User Reactions
Meta's own user research:
- Ads labeled AI: 30% negative reaction (some pushback)
- Undisclosed AI ads, caught later: 70%+ negative reaction (trust collapses)
Implication: Disclosing AI openly pays off long-term. It protects brand trust better than hiding it.
So What Do We Do?
Response by scale:
Small brands:
- Always check the box when uploading AI creatives
- Never composite celebrity faces with AI
- Stick to virtual characters and product-centric AI use
Mid-tier:
- Document your AI usage guidelines
- Train team members and freelancers on the policy
- Review the latest policy quarterly
Large scale:
- Legal review process
- AI usage logging
- C2PA metadata management
Where This Is Heading
Short term (2025-2026):
- Meta's AI detection keeps getting more accurate
- Stronger user-facing label UI
- Harsher penalties for repeat violators
Long term (2026+):
- EU AI Act and similar regulations will tighten Meta's policies further
- Possible standardization of auto-watermarking for all AI content
- AI transparency becomes a competitive advantage for advertisers
Ad policy, review, and compliance are covered in Meta Ads Book 1.