I have not researched closely enough but I think it falls under what corporations do. They are commercially licensed, you cannot use them freely, and crucially they were trained using data scraped at an industrial scale, contributing to degradation of the Web for humans.