[Quoth]

Originally Posted by salamanderjuice

I mean, they do actually say what they are fed and how it's curated:

https://arxiv.org/pdf/2302.13971
https://arxiv.org/pdf/2505.09388

In general they don't and can you believe them? MS and Google have not been great for truthiness and transparency.

However it make no difference. It's only any use if you can separately verify it. It's mostly plausible only if you are not already an expert.