@gamingonlinux For what it's worth, it's because the datasets they use will inevitably contain copyrighted work (articles, posts, websites, you get the idea). I don't think OpenAI are specifically targeting copyrighted works, but the content of the datasets being used makes copyright infringement inevitable.
There's an entire ethical conversation to be had with this sort of mass-scale scraping that goes beyond machine learning research.
Like I've been saying for many, many years now: If a company can't operate while obeying the law then they shouldn't operate at all. No corporation has a right to exist, if Facebook can't run their business if forced to moderate their platform, then they should shut down, if OpenAI can't operate without violating copyright then they should shut down.
@gamingonlinux Well, but it is fairly possible to pay the authors for using their work, isn't it? Is also very easy to ask for an permission to use someone's work, especially commercially.
So why the fuck they don't do that?!
How was that?
You're taking my stuff or using my work = you're paying my bills!
Problem solved and case closed.
But if not, we will see each other very soon in the court.
Simple as fuck!
@gamingonlinux Companies simultaneously be like "you're infringing my copyright by playing my 30-year-old game for free" and "it would be impossible for us to make a profit without infringing your copyright"