@trophaeum
When you train an AI by scraping the internet, you are making an implicit philosophical claim: that knowledge is a resource to be extracted, that the world is a mine of information waiting to be processed, that the relationship between the AI and the knowledge it holds is one of ownership and accumulation.
This is not a neutral assumption. It is a particular view of reality ; one that most of the AI industry has inherited from the economic systems that fund it, without ever examining whether it is the right view for the purpose at hand.