Unverified50% confidenceFactExact time
When the semantic direction of text prompts conflicts with the motion characteristics of a reference video, the model must negotiate between two signals, often resulting in unstable motion output.
1
Sources
50%
Confidence
Long-term
Relevance
8/19/2026
First Seen
Sources
AI视频生成角色动作太夸张？实用技巧驯服过度活跃的动作
redditr/comfyui8/18/2026
Related Claims
UnverifiedOne of the prominent pain points of current text-to-video models is the uncontrollability and inconsistency of motion, where the same prompt may generate completely different actions.69% similarUnverifiedAdding keywords describing slow, smooth, and natural movement to prompts (such as "slow, deliberate movements" and "subtle gestures") can reduce the problem of AI video characters moving too actively.64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/775720API
curl https://kongchang.com/api/v1/knowledge/claims/775720MCP
get_claim(id=775720)