Meta on July 7, 2026 released Muse Image, the first image generation model built by Meta Superintelligence Labs. A feature that generates images by referencing photos from public Instagram accounts is enabled by default, and it has drawn a privacy backlash almost immediately after launch.
July 7, 2026 · Meta Superintelligence Labs
Meta's Muse Image Launches — and Instantly Sparks a Privacy Backlash
Meta's first in-house image model can generate art by referencing photos from any public Instagram account via a simple @-mention. The catch: the feature is on by default , with no notice or explicit consent.
#2
Ranked 2nd on Arena Elo at launch — for text-to-image, single- and multi-image editing
@
One @-mention pulls in a public account's photos — no notice, no explicit consent
ON
Default setting for all public accounts; opt-out buried in settings
The privacy gap: consent vs. capability
To disable AI reuse of your public photos, users must navigate three layers deep.
Profile
→
Menu
→
"Sharing and reuse"
→
Turn AI reuse OFF
Previously generated images are not deleted , and some users report the opt-out was not yet applied during rollout.
Praised by creators
High fidelity to complex prompts
Naturally blends multiple reference images
Precise editing via sketch / annotation
Generates working QR codes, charts & GIFs
Under scrutiny
Default-on reuse of public photos for AI
Likeness use without consent
Deepfake risk raised by privacy groups
Opt-out guides spreading rapidly online
"Consent, not capability."
Muse Image works with the Muse Spark foundation model for agentic generation — search, code execution and self-improvement — as Meta shifts from third-party models to its own. But the near-term battle is over defaults, discoverability, and the handling of images already made.
Continue reading The rest of this article is for AI News Blitz readers. Choose an option below to keep reading.
Already purchased? Sign in ✓ Signed in — this article isn’t included in your current plan.Unlocking the full article…