Arguments for Rigorous AI Training Data Transparency (read.misalignedmag.com)

What do “frontier AI labs” actually train their models on? Do we know? Do they even know? There are arguments to be made we should all know, and in detail.