>But AI developers/companies seem to almost universally believe they have fair-use rights to train their models on any data they can get their hands on
Google and many other companies for almost 2 decades have spent their time scraping petabytes of data from the web. A lot of that with no expectation of payment. Some companies became billionaires off of that ability to freely access mass bulks of data.
Data scraping has always been a grey area, but I find it strange how it's suddenly taken a turn for some people whenever modern AI comes up. We can't really be drawing lines based on what we feel is good/evil, because we will never agree as a whole on what is good/evil.
Google and many other companies for almost 2 decades have spent their time scraping petabytes of data from the web. A lot of that with no expectation of payment. Some companies became billionaires off of that ability to freely access mass bulks of data.
Data scraping has always been a grey area, but I find it strange how it's suddenly taken a turn for some people whenever modern AI comes up. We can't really be drawing lines based on what we feel is good/evil, because we will never agree as a whole on what is good/evil.