The new process makes use of Meta’s Muse Image model, which was developed by the company’s next-level Superintelligence Labs AI research group. The updated AI image generation tools will expand marketers’ capacity to restyle ad images, inspired by existing ads. The model will also let marketers generate still images from video assets.
This audio is auto-generated. Please let us know if you have feedback.
The new process makes use of Meta’s Muse Image model, which was developed by the company’s next-level Superintelligence Labs AI research group. The updated AI image generation tools will expand marketers’ capacity to restyle ad images, inspired by existing ads. The model will also let marketers generate still images from video assets.
As explained by Meta: “Muse Image is built with agentic visual reasoning and self-refinement, meaning it understands complex creative briefs the way a designer would — not just individual keywords. Muse Image brings native reasoning to the creative process to adjust elements, swap styles, and create variations based on the advertiser’s creative, resulting in high-quality, on-brand ad variations with fewer iterations.”
Aside from utilizing Meta’s most advanced AI systems, the tool is also based on Meta’s systematic understanding of promotions, which will enable it to create content that better aligns with what’s generating response from each app’s audience.
Meta said that advertisers who’ve used its AI tools to generate multiple variations of their promotions thus far have seen strong results. Early testers of Muse Image have said that the system generates higher-quality creative outputs, including improved photorealism and the ability to preserve product integrity.
Muse Image will also enable expanded generation capacity.
“Meta AI can now reimagine a shopper’s room, powered by Muse Image’s photorealistic image generation,” Meta said. “When someone uploads a photo of their space, Meta AI restyles it with real products from businesses’ product catalogs — showing how each item looks in their actual room. They can compare options, keep the conversation going to refine the look, or tap through to shop directly on a brand’s website.”
This expanded visual discovery process is the next step in Meta’s AI ad plan, by offering new ways for users to interact and engage with promotions.
Meta is heavily invested in AI infrastructure, and it’s now looking for ways to recoup those costs, possibly via a range of AI innovations.
Muse Image is its latest project on this front as Meta’s Superintelligence team works to build better tools for a broader range of customers.
As such, Meta has a lot riding on the success of these models. That means it may be worth exploring its generative AI options.
Meta said that more than 8 million advertisers are already using at least one of its generative AI ad creative tools.
