The billion-dollar lawsuit brought by The New York Times against OpenAI et al. has raised many important issues and gotten a lot of attention in general.
One position from OpenAI is comment-worthy, but I don’t see receiving attention.
OpenAI contends that AI progress is impossible without using copyrighted material and that that justifies their use of it.
I find this position to be ludicrous for two reasons.
First, it can’t possibly be true. Anything that has been created by others under copyright could be created by a team created by OpenAI.
Second, regardless of whether it is true, an ingredient being necessary is not an excuse for not paying for it. If you or I started a business that was only viable because it stole something (or perhaps failed to pay its workers a living wage), that would not be a good reason to do illegal or unethical things; it would be a reason why that business wasn’t viable and shouldn’t exist in that form.
Whether it is creating their own writing teams or properly buying appropriate rights, it’s hard for me to imagine that a multi-billion dollar high-flying can’t afford it.
I find the position to be morally bankrupt.
That said, it may well be that the system could well be built from freely available text.
And it may be that this counts as fair use.
I am not weighing in on either the factual or legal questions.
Only the moral one: “It’s necessary for my business or even my entire industry,” will never be a valid justification for theft.
