Does traning AI/ML-models on AI-generated content causes collapse on the quality of the output?
Does traning AI/ML-models on AI-generated content causes collapse on the quality of the output?
If so are these programs that claim to 'poison' the training datasets effective ?