Meta wants advertisers to use its AI tools to boost performance. But the tools are buggy and sometimes change product images in unwanted ways. That leaves brands doing extra work to fix the mistakes.
Last month, outdoor retailer REI experienced a high-profile incident where Meta's AI altered ad images. A small business called Quite Literally Books had its Valentine's Day ad changed without permission. Photographer and marketer Abigail Hogue saw her work modified by the machine.
Ads consultant Jessica Gleim works with female-founded brands. She said the tools are not usable. "It's not usable to help my clients grow their business," Gleim told Business Insider.
Robert Webster, who leads TAU Marketing - a firm overseeing around $500 million annually in advertising across various platforms - cautioned that brands could suffer reputational harm. "When the AI starts generating weird creative or making unapproved changes, it can quietly damage brand perception - especially for anyone who cares about consistency," he said.
Luke Jonas, chief growth officer at Nest Commerce, offered a visual analogy: "A machine optimizing for 6 million advertisers will occasionally give you two handlebars." The AI creates strange images, like a product with two handlebars instead of one.
Get the market news that matters in a five-minute read with Market Briefs, our free daily newsletter
Karissa Tuccio, the executive director of social and influencer for Mediassociates, noted that 15 of her clients encountered the issue on a regular basis. Last week - relative to the article's July 13, 2026 publication date - she reported the bug still existed. "So that leads me to believe it has not been resolved," Tuccio said.
Rok Hladnik, CEO of Flat Circle - an agency that handles around $200 million in yearly Meta ad investments - stated his team now treats the additional manual labor as standard procedure. "We somehow accepted that as a new standard operating procedure," he said.
Meta's official response: The company said "AI can make mistakes and that it is the advertiser's responsibility to review the AI outputs." Last month, Meta began automatically adding "AI info" labels to ads that use the tools. Last week, it started rolling out a new AI image model called Muse Image.
Meta also developed an internal quality-control dashboard for large advertisers to check that AI enhancements are turned off. But advertisers remain skeptical.
Danny Weisman, cofounder of Obsessed Media, downplayed the errors. "It's not like someone's hand is missing," he said.
The issues underscore a growing challenge for Meta as it pushes AI automation to compete with rivals like Google and TikTok. While AI can optimize ad performance at scale, the glitches force advertisers to treat automated tools with caution, often requiring additional layers of review that offset the promised efficiency gains. For small businesses especially, the extra manual work can strain limited resources, making Meta's tools less appealing.
Advertisers are watching whether Meta fixes the bug that automatically enables AI settings and whether the new Muse Image model reduces or increases the problems.
Join Market Briefs, our free daily newsletter, for a quick daily rundown of the markets
