I think AI models are different form ML models, atleast mathematically. ML models and the mathematics behind it are explained thoroughly and people (who learnt about ML models) why certain ML models work or do not work. The AI models which are sprining left and right now I doubt people(again who learnt about AI models) have complete understanding about why they work, similar to how X-rays were used by doctors but they did not know why it worked. I guess for buisness case of ChatGPT if it is brining in the money why should it not be used?
I don't think that's a valid distinction. There's no such thing as "AI" models imho. ChatGPT is also just ML. It's just a specific type of model that's hard to explain because it has so many parameters and embeddings in-between layers and stuff like that. It's not as easily explainable as decision tree-based models maybe. You would be surprised by how many models are actually hard to explain.
