AI Definitions: Model Collapse

AI model collapse - The idea that AI can eat itself by running out of fresh data, so that it begins to train on it’s on product or the product of another AI. This would magnify errors and bias and make rare data more likely to be lost.

More AI definitions here.