Meta launched its first in-house AI image generator, Muse Image, introducing a controversial feature that allows tagging public accounts.
Tech giant Meta has officially thrown its hat into the native text-to-image race by launching its very first in-house artificial intelligence image generator. Developed by the company’s specialized Superintelligence Labs division, the new model, officially dubbed “Muse Image”, began its public rollout on Tuesday, July 7, 2026. The advanced tool is being embedded directly into the Meta AI standalone application, WhatsApp direct messages, and Instagram Stories for users located within the United States. Meta’s primary goal with this launch is to completely replace its reliance on third-party image software providers like Midjourney, court high-spending digital advertisers, and keep users deeply engaged inside its massive social ecosystem by making AI image generation a highly personalized, daily routine.
The technology behind Muse Image marks a significant shift from traditional AI art tools. According to Meta’s Chief AI Officer Alexandr Wang, Muse Image is engineered to be “agentic,” meaning it works hand-in-hand with the company’s previously released Muse Spark language model to logically plan layouts, run web searches, and self-refine its own drawings before showing them to the user. This advanced capability allows the model to successfully handle complex tasks that have famously confused older AI tools for years, such as embedding completely readable text into a graphic, generating working QR codes, or seamlessly blending multiple real photos into a single, cohesive image. Users can also utilize a new interactive editor to draw or circle specific parts of an existing photograph, allowing them to tweak small details rather than forcing the AI to rewrite and regenerate the entire picture from scratch.
However, the launch has instantly sparked an intense global conversation regarding user privacy due to a highly unique and controversial feature. Inside the Meta AI app, users can physically “@-mention” or tag any public Instagram account within a prompt, instructing the AI to instantly pull public photos from that profile and insert that specific person’s likeness into a freshly generated image. While Meta frames this as an entertaining way for friends to create custom digital postcards or place themselves into whimsical video game styles together, privacy advocates have sounded immediate alarms. Critics point out that allowing the tool to pull real people into AI-altered photographs without their explicit, case-by-case consent creates a massive digital safety hazard.
See Also: Vercel CEO Outlines the High-Stakes Battle to Separate AI Models From Smart Agents
Compounding the controversy is how Meta decided to roll out the framework. The profile-tagging feature is enabled by default for every single public account on Instagram. While Meta points out that users can manually enter their account privacy settings to opt out of the AI remixing feature, or simply flip their entire profile to a private setting to block future generations, independent tech researchers argue that a default-on approach leaves millions of everyday users vulnerable to unauthorized likeness replication.
Despite the looming privacy debates, Meta is moving full steam ahead with plans to monetize the technology. While everyday casual use of Muse Image will remain entirely free up to a designated monthly limit, heavy content creators who exhaust their credits will be prompted to upgrade to a paid “Meta One” monthly subscription. Furthermore, Meta intends to roll out Muse Image to Facebook and Messenger later this year while integrating the tool directly into its Advantage Plus marketing platform. This will allow corporate advertisers and creative agencies to automatically generate hundreds of distinct, on-brand image variants for social media campaigns with a single click, signaling a brand new chapter for automated digital marketing.
