
@MrLukowski Because it doesn't know or understand anything. It's all based on statistical analysis and what things are likely to go together. It has no concept of "teeth" or "mouths", only that images containing mouth-like features, are likely to have white objects in them
English






