It appears that at least one company loudly trumpetting the wonder of its artificial intelligence systems may have trained those systems using pirated copyright material.
The excuse seems somewhere between "we need it because otherwise we couldn't make it work quickly, easily and cheaply" and "what are you going to do about it?".
It's difficult to argue with such genius logic. Starting with the second point, they are probably correct that any affected authors, even if they all worked together, would be unlikely to afford to spend the money, time and even more money on a legal case without end.
On the one side, a large corporation with its in-house legal department (which, one assumes, it pays anyway), whose dear lawyers can beaver away producing hundreds, maybe thousands, of pages of obfuscation and delay, perhaps using AI to assist.
On the other side, a rapidly escalating legal bill for a bunch of cash-strapped authors. It's not really going to happen, is it?
But mockery is cheap and, even if not effective, can provide a small respite against the feeling of helplessness.
So here's my generous offer which I am making to these wondrous technology overlords.
I'd like to build a social network instant messaging artificial intelligence systemy thingy. But the only way I can make it work quickly, easily and cheaply is if I copy a load of your code. Fortunately, I've managed to find a repository of the entirety of your codebase so, given your heavily publicised views on the use of such material as a mighty enabler, I shall use as much of it as I see fit.
In return, you're welcome to feed my books into your AI meat-grinder. Deal?
(Please note -- this is, of course, do I really have to spell it out, a whimsical piece of satirical sarcasm. Nothing has been stolen, pirated or otherwise inappropriately acquired by me. Would that the AI companies could say the same thing.)
