If your Meta ad account just got restricted — or your ROAS tanked overnight for no apparent reason — there’s a good chance Meta’s own AI did it to you.
Here’s the hard truth: Meta has been silently enabling AI-powered creative features inside ad accounts without advertiser consent. These features modify your ad copy, swap your images, generate new creative variations, and overlay music on your videos — all automatically. And when those AI-generated modifications trip Meta’s own policy enforcement system, your account takes the hit. Not Meta’s AI. You.
This isn’t a hypothetical. We’re seeing it across accounts we manage, and the advertiser forums are full of brands dealing with the same thing. Let’s break down exactly what’s happening, why it’s dangerous, and how to protect your account.
What Meta’s AI Is Actually Doing Inside Your Account
Meta has rolled out several AI creative generation features over the past 12 months. The problem isn’t that these tools exist — it’s that Meta auto-enables them without clearly telling you, and the output isn’t always compliant with Meta’s own advertising policies.
Here are the three areas where AI is most likely modifying your ads right now:
1. Advantage+ Creative Enhancements
At the ad level, Meta’s “Advantage+ Creative Enhancements” section can contain up to 10 different AI toggles. These include:
- Text improvements: Meta rewrites your primary text, headline, and description
- Image enhancements: Auto-adjusts brightness, contrast, and applies filters
- Music overlay: Adds background music to your video ads
- Visual touch-ups: Expands or crops images, adds text overlays
- AI-generated backgrounds: Replaces product backgrounds in catalog ads
The issue? Meta adds new toggles over time without notifying you. An ad you set up clean three months ago might now have two or three new AI features enabled that you never approved.
2. Account-Level “Creative Features” Testing
In your Advertising Settings (the gear icon in Ads Manager), there’s a “Creative Features” module. Inside it, you’ll find two toggles:
- Test New Creative Features
- Test New AI Generation Features
Meta periodically re-enables these — sometimes within days of you turning them off. One media buyer reported finding them re-enabled three separate times in a single week. When these are on, Meta is using your budget and your brand to test their experimental AI tools.
3. Catalog-Level Background Generation
If you run catalog or shopping ads, Meta’s Commerce Manager can auto-generate new product image backgrounds. These AI-generated backgrounds can misrepresent your product, create misleading context, or produce imagery that doesn’t match your brand guidelines.
Why This Gets Accounts Shut Down
Here’s where it gets ugly. Meta’s AI generates creative variations → those variations get served to users → Meta’s separate policy enforcement AI reviews the served ads → and flags policy violations. The left hand doesn’t know what the right hand is doing.
The most common violations we’re seeing from AI-generated creative modifications:
- Before/after claims in health and wellness: The AI rewrites body copy to include transformation language that violates Meta’s health advertising policies
- Misleading claims: AI-generated text makes promises your original copy carefully avoided
- Personal attribute violations: The AI inserts language that asserts or implies things about the viewer (“Are you struggling with…?”) which violates Meta’s discriminatory practices policies
- Brand safety issues: Music overlays or background modifications create inappropriate brand associations
- Copyright concerns: AI-generated backgrounds may inadvertently reproduce copyrighted elements
The result? Your ad gets disapproved. Enough disapprovals → your account gets flagged for “repeated policy violations.” Enough flags → restricted or disabled. And at no point in this chain does Meta acknowledge that their AI created the violation.
The B2B Problem: It’s Even Worse for Regulated Industries
If you’re in supplements, health, finance, alcohol, or any regulated category, the risk multiplies. These verticals have strict advertising guidelines, and Meta’s AI has zero awareness of your specific compliance requirements.
We’ve seen cases where:
- A supplement brand’s carefully compliant copy was rewritten by Meta AI to include efficacy claims that violate FDA advertising guidelines
- A financial services company had AI-generated songs added to their professional B2B video ads
- A skincare brand had AI-modified images that created implied before/after comparisons
These aren’t minor inconveniences. For brands in regulated industries, a single AI-generated policy violation can trigger a manual review of your entire account history.
How to Protect Your Account: The Full Checklist
This isn’t a “check once and forget it” situation. Meta re-enables these features regularly. Build this into your weekly account maintenance.
Step 1: Disable Account-Level AI Testing
- Open Meta Ads Manager
- Click the ⚙️ gear icon → Advertising Settings
- Find the “Creative Features” module
- Turn OFF: “Test New Creative Features”
- Turn OFF: “Test New AI Generation Features”
Check this weekly. Meta has a pattern of re-enabling these silently.
Step 2: Audit Every Active Ad’s Creative Enhancements
- Open each active ad in your account
- Scroll past the CTA section to the “Advantage+ Creative” area
- Review every toggle — some ads will show 4 options, others up to 10
- Disable anything you didn’t explicitly enable
- Pay special attention to: Text improvements, Music, Visual touch-ups
New toggles appear over time without notification. An ad that had 4 enhancement options when you created it might now have 7.
Step 3: Lock Down Catalog AI Features
- Go to Meta Commerce Manager
- Check individual products for a “Backgrounds” section
- Toggle OFF background generation for each product
- Review your catalog feed for any AI-modified images
Step 4: Implement a Weekly AI Audit
Add this to your weekly account maintenance checklist:
- ☐ Check Advertising Settings → Creative Features (both toggles OFF)
- ☐ Spot-check 3-5 active ads for new Advantage+ Creative toggles
- ☐ Review any recent ad disapprovals for AI-modified content
- ☐ Check Commerce Manager for catalog background generation
- ☐ Review ad previews on mobile to confirm what’s actually being served
Step 5: Document Everything
If your account does get flagged or restricted, you’ll need evidence that the violation was caused by Meta’s AI, not your team. Screenshot your original creative, your ad settings, and the AI modifications. This is critical for appeals.
Why Meta Keeps Pushing This
Let’s be real about what’s happening here. Meta’s AI creative tools serve Meta’s interests — not yours. More creative variations = more auction entries = more ad spend. Meta’s recommendation engine is optimized to increase your spend, not your ROAS.
Every “recommendation” in Ads Manager — turn on Advantage+ audience, enable creative enhancements, expand your placements — is designed to give Meta more control over your budget allocation. Some of these genuinely help. But when Meta auto-enables experimental AI features using your brand and your budget as the test subject, that’s not optimization. That’s a conflict of interest.
The brands that perform best on Meta in 2026 are the ones that maintain tight control over their creative, their targeting, and their account settings — while using Meta’s AI selectively and intentionally, not by default.
The Bottom Line
Meta’s AI creative tools aren’t inherently bad. Some of them, used deliberately, can improve performance. The problem is the auto-enrollment, the silent re-enabling, and the complete lack of accountability when AI-generated content triggers policy violations.
If you’re running ecommerce ads on Meta, this isn’t optional maintenance. It’s account survival. One AI-generated policy violation you didn’t create can cascade into a restricted account — and once Meta’s enforcement system labels you a repeat offender, the appeals process is brutal.
Check your settings today. Not tomorrow. Today.
And if you’re spending serious budget on Meta and want someone to audit your account settings, creative controls, and policy exposure — book a strategy session. We’ll walk through your account and show you exactly where you’re exposed.
