Meta's AI ad tools are quietly changing your campaigns without asking. A new investigation published July 14, 2026, by Business Insider and TheNextWeb documented eight advertisers and agency executives who all confirmed that cleaning up after Meta AI has become routine work inside their organizations. For any ecommerce brand running paid media on Meta, this is not a hypothetical future risk. It is happening to campaigns right now, and the settings you turned off may already be back on.
Meta's Advantage+ Creative suite is the central vehicle for this automation push. It includes tools that adjust images, add overlays, alter aspect ratios, generate backgrounds, and in some cases replace or significantly alter the creative you uploaded. Meta describes it as a way to improve performance by showing audiences the version of an ad that is most likely to get a result.
The problem documented in the investigation is not that AI optimization exists. It is that brands are being enrolled in these features without explicit consent, and the results are producing the exact opposite of brand-safe advertising.
REI, the outdoor retail brand, told reporters that Meta auto-enrolled them in an AI feature that produced what REI called "inaccurate" images. REI had not opted in. The investigation also documented a pyjama dress turned into a separate shirt and trousers, men added to a creative designed for a women's networking group, and a bicycle generated for REI with two handlebars instead of one.
"Cleaning up after Meta's AI has become routine." — Confirmed by all eight advertisers and agency executives interviewed in the Business Insider investigation
One marketer described shooting an entire campaign herself, watching Meta's AI garble the text and warp her products, and then having friends message her to say it looked like "AI slop." She requested a refund. Weeks later it had not arrived.
This is the harder question, and the one most brands are not set up to answer well. Meta does not send an alert when Advantage+ Creative modifies a piece of creative. The altered version simply runs. If you are not actively comparing your source files against what is live in Ads Manager, you may not notice until a customer sends a screenshot or your brand team flags something in the wild.
The investigation surfaced a particularly alarming bug report: multiple agencies confirmed that after turning AI settings off inside Ads Manager, the settings toggled back on. Meta's own representative reportedly offered to manually check specific ad IDs before a launch, which suggests the bug was not yet resolved at the time of reporting. You cannot simply turn something off once and trust it stayed off.
Meta's reach is not a small variable here. The platform reaches 3.5 billion people daily and earned $196 billion in ad revenue in 2025. Millions of advertisers are active across its systems. When a default setting silently modifies creative, it modifies it at scale.
Meta's official response: "AI can make mistakes and that it is the advertiser's responsibility to review the AI outputs." The burden of proof falls on you, not the platform.
Meta's official position is that AI can make mistakes and that it is the advertiser's responsibility to review AI outputs before campaigns run. That framing places the quality control burden entirely on brands. Given that some settings turn themselves back on after being disabled, passive trust is not a viable strategy.
Here is what we are advising every ecommerce brand to do immediately:
The framing of "AI can make mistakes" implies an occasional glitch. What the investigation actually describes is a systemic pattern: auto-enrollment without consent, creative modification at scale, settings that reassert themselves, and no proactive notification to brands when their ads are being altered.
For 7-figure and 8-figure ecommerce brands where product photography, brand voice, and visual consistency represent real investment, this is not a minor inconvenience. An altered image that reaches millions of people before anyone catches it is a brand event. A garbled campaign for a product you spent months developing is a real revenue and trust problem.
Meta's Advantage+ tools do have legitimate applications. Automated placement optimization and audience broadening have shown measurable performance lift for many accounts. The issue is not AI-assisted advertising as a category. The issue is default enrollment in creative modification without consent, and a feedback loop that does not require Meta to fix errors before brands absorb the cost.
Brands that are running any volume of Meta spend should treat this as a standing operational checklist item, not a one-time audit.
CTC helps 7-figure and 8-figure ecommerce brands build Meta advertising systems with the right human oversight, creative governance, and measurement to protect and grow revenue.
Common Thread Collective is the leading source of strategy and insight serving DTC ecommerce businesses. From agency services to educational resources for eccomerce leaders and marketers, CTC is committed to helping you do your job better.
For more content like this, sign up for our newsletter, listen to our podcast, or follow us on YouTube or Twitter.