Unverified50% confidenceFactExact time
AI模型缺乏具身经验，只能从像素级统计共现中学习动作表征，因此在生成涉及复杂肢体交互的图像时常出现腿部结构错误、动作方向反常等问题
1
Sources
50%
Confidence
Long-term
Relevance
7/20/2026
First Seen
Sources
Related Claims
VerifiedTypical defects in AI-generated video include facial distortion (face merging), abnormal finger counts, limb clipping, objects disappearing or deforming between frames, and physically implausible motion.77% similarUnverifiedAI-generated motion code may contain physically implausible situations (such as an arm passing through the body or a joint bending in reverse), requiring human review.76% similarUnverifiedAI image models are fundamentally learning visual statistical correlations rather than understanding physical laws or ergonomic logic.73% similarVerifiedThere are generally two implementation paths for AI models to internalize physical laws: incorporating large amounts of synthetic videos generated by physics simulation engines for statistical learning, or introducing explicit physics constraint modules into the model architecture.69% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/570077API
curl https://kongchang.com/api/v1/knowledge/claims/570077MCP
get_claim(id=570077)