Meta is pushing advertisers to hand their campaigns to its AI. A Business Insider investigation found the tools spitting out gibberish copy, mangled limbs and products that no longer look like the products. Meta’s reply, in effect: that’s your problem.
Meta wants brands to let its AI build their ads. A Business Insider investigation shows what that looks like in practice, and it is not pretty.
Business Insider spoke with eight advertisers and agency executives. All said that cleaning up after Meta’s AI had become routine. The tools, they said, are clunky, and they generate misrepresentations and outright absurdities.
Twisted limbs, garbled text, the wrong product
The examples are vivid. Meta suggested turning a client’s pyjama dress into a shirt and trousers, ads consultant Jessica Gleim told BI. For a women’s networking group in Montana, Meta’s bright idea was to add men.
The 💜 of EU tech
The latest rumblings from the EU tech scene, a story from our wise ol’ founder Boris, and some questionable AI art. It’s free, every week, in your inbox. Sign up now!
The outdoor retailer REI drew a backlash last month over an Instagram ad showing a nonsensical bike with two handlebars. REI said Meta had “auto-enrolled” it in an AI feature that produced an “inaccurate” image.
Around Valentine’s Day, marketer Abigail Hogue watched Meta’s AI garble the text and warp the products in a campaign she had shot herself. Friends messaged to accuse her of posting “AI slop”. She asked for a refund. Weeks later, she said, it still had not arrived.
‘That’s on you’
Meta’s response puts the problem back on the advertiser. Its terms say “AI can make mistakes and that it is the advertiser’s responsibility to review the AI outputs,” a spokesperson told BI.
For a brand, that is not a quirky demo. An ad is a legal and commercial artefact. When an automated tool rewrites your copy or alters your product, you own claims you never made, running under your name and paid for with your budget.
It gets worse when the tools switch on by themselves. Several agencies told BI of a bug that toggled AI settings back on. One said a Meta rep offered to check her ad IDs by hand before a big launch, which rather suggested the bug was still live.
Why Meta keeps pushing
The pressure is deliberate. Meta’s advertising business pulled in around $196bn last year, and it reaches 3.5 billion people a day. Its Advantage+ tools are meant to turn that scale into automated, higher-performing ads. Meta says millions of advertisers are finding value in them.
The same urgency has dogged Meta’s consumer AI. It pulled a feature from its new Muse Image model days after launch, after a backlash. Its own AI-image detector then missed most fakes once they were cropped.
An industry-wide mess
Meta is not alone. Google now labels AI-made ads, and its Performance Max tool also crops and rewrites creative automatically. But brands told BI it has largely avoided Meta’s worst failures. Reddit, meanwhile, is fighting AI marketing “slop” with more AI.
The trap is that Meta works. It has the reach, the data and the results, so most brands cannot walk away. That is exactly why it can keep pushing AI at them, and tell the ones who complain that the mess is their own fault.
