Google Adds AI Video Remixing to YouTube Shorts via Gemini Omni

Bottom Line

According to The Verge, Google has integrated Gemini Omni into YouTube Shorts as a Remix feature, letting users both restyle video clips and edit their contents—transforming footage into anime or inserting themselves into others’ videos. Creators retain control through a disable toggle, and remixed videos carry watermarks and attribution links to protect original creators.

Gemini Omni Remix Capabilities

The new Remix option appears in the bottom menu of any YouTube Short. When a user selects it, Gemini Omni offers two classes of edits: stylistic transformations and content manipulation. According to The Verge, stylistic options include converting a video to pixel art, anime, or found-footage horror film aesthetics. Beyond restyling, the feature also permits users to alter video contents—enlarging heads, inserting background actors, dressing people in different costumes, or compositing the remixing user directly into the original clip.

Creator Control and Attribution

Google has built in creator protections via an opt-out mechanism. According to The Verge, content creators uploading Shorts can disable remixing on their uploads, allowing them to prevent manipulation of their videos entirely. This is particularly relevant for creators uploading sensitive content—such as footage of children—where allowing algorithmic or user-driven edits would be inappropriate.

To preserve attribution, remixed Shorts automatically include a digital watermark and a link back to the original video. This mechanism acknowledges the source creator and directs engagement back to the original upload, mitigating some of the friction between remix culture and creator consent.

Why This Matters

This feature represents a significant expansion of generative AI’s footprint in short-form video, the dominant format for creator-driven platforms. The combination of stylistic remix (similar to existing filter-based editing) with generative content insertion (faces, actors, objects) creates new possibilities for collaborative and transformative video creation—but also raises questions about consent and deepfake potential at scale.

The opt-out model places the burden of protection on creators rather than requiring explicit consent for remixing. While Google’s watermarking addresses attribution, the ability for any user to insert themselves into others’ videos or manipulate people’s appearances could strain creator relationships if remixes go viral in unintended contexts. Industry norms around consent for generative remixing remain unsettled; this launch may become a test case for how platforms balance creation tools with creator agency.