Google demonstrates multi-turn conversational editing for video generation
Refine the same clip turn by turn in conversation without starting over.
AI BriefingGoogle demonstrated that video generation can progressively modify the same clip through multiple rounds of conversation, without starting over. In the example, Turn 1 generates an octopus with a rippling effect, while the follow-up prompt in Turn 2 changes the protagonist into a dancer, with the skirt precisely reproducing the original movement.
Source: Google · x.com
Event follow-up1
View all event coverage- Google releases Gemini Omni 1.2 Flash: 15-second video generation, with a 30% price cut to $0.07/secondGoogle Blog:AI · 11 days ago