Meta announced on Tuesday the Muse Image AI image generator and a forthcoming Muse Video product. Muse Image can use content from public Instagram accounts as input when generating new images.
A hands-on test: fabricating realistic photos in minutes
A Techtud reporter logged into the Meta AI web app (meta.ai) from a desktop using their Instagram account and asked the chatbot to create a photo of them skydiving with Daniel Craig. Within about a minute the system produced an image showing the reporter and Daniel Craig laughing while falling from an altitude of 3,000 meters.
The reporter then selected additional options (for example, mountains in the background, sunset and an action-camera blur), and the image was adjusted accordingly. The example demonstrates how a few clicks can generate images that appear real despite never having been actually taken.
Privacy concern: images of people can be generated without consent
The article highlights that Muse Image is capable of producing images based on data from other people's Instagram accounts. This applies even to individuals the user never followed or knew on Instagram, and it can occur without the subject's permission.
Public Instagram account settings allow others to access and reuse posted photos, which underlies this capability. Instagram updated its relevant guidance stating: “If your account is public, other Instagram users may create new Reels, posts, or Stories that reuse part or all of your posted photos, videos, or Reels. In addition, others may use Meta’s AI-powered features to create content using your Instagram content. Depending on the other user’s settings, this may mean your reused content could appear in search engine results.”
Meta's response: built-in safeguards and enforcement
Responding to a Gizmodo article, a Meta spokesperson said that Muse Image includes built-in protections intended to prevent creation of content that violates policy. Accordingly, the system is designed not to produce violent, sexual, or malicious images of real people. Meta also said that policy-violating content will be subject to enforcement actions whether it is reported by users or detected by automated systems.
Why this matters
The story underscores privacy and misuse risks arising where generative AI systems access publicly available social media material. While Meta describes safeguards to limit harmful outputs, the deployment of tools that can fabricate realistic images from public posts raises practical questions about how effectively these protections work and how individuals' rights are preserved when public content can be repurposed without consent.
