Meta introduced Muse Image and offered an early look at Muse Video, expanding its push into AI-generated media across its consumer apps. Muse Image operates as an “agent,” invoking search and coding tools, iterating on its own outputs, and scaling quality with more test-time compute; it also integrates with Muse Spark for joint planning and tool sharing. The model supports precise editing, multi-reference composition and ranks near the top of human-preference leaderboards, while a new invisible watermark, Content Seal, aims to help verify AI-made images. Muse Video, built on the same pretraining base, promises strong visual fidelity and native audio, though Meta acknowledges challenges with audio sync and fast motion. Muse Image rolls out across the Meta AI app, meta.ai, Instagram Stories in the U.S., and limited WhatsApp markets, with Facebook to follow; Muse Video is slated “coming soon” to creators. The launch underscores intensifying competition in AI media generation as platforms race to blend creation tools with social and commercial use cases.
Related article:
Content Credentials: C2PA’s standard for provenance and attribution
