I increasingly think OpenAI is trying to destroy the concept of copyright through desensitization.
Copyright is a huge threat to their profits. They are facing multiple massive copyright lawsuits.
So they release Sora, knowing it can output copyrighted IP and knowing people will use it for this. Ghibli was a trial run.
It floods the internet. More importantly, other apps will follow their lead. I suspect we will see this soon - it’s what happened with ChatGPT and training, everyone thought ‘if they can get away with it, perhaps we can too’.
So it becomes the new norm, used by millions, and it’s that much harder for society to backtrack. Lawsuits are expensive to bring. It’s harder to defend your rights when everyone is exploiting your IP.
They are trying to shift what we consider normal, so that copyright stops being an issue for them.
We should all be trying to ensure this strategy does not succeed.
It should not be controversial to believe both of these statements:
1. AI is capable & can be useful
2. The way AI is currently built is unfair and probably illegal
AI boosters often say licensing training data shouldn’t be required because only big companies could afford the licensing fees, so it would kill competition. This is not a good argument.
If you want to build a business that uses copyrighted work commercially, you have to work out a way to pay for it. This applies to businesses from book publishers to streaming services. By the boosters’ logic, Penguin Random House and Spotify should also not be required to pay for the content they use. After all, licensing all those works costs money, and these costs make it hard for smaller companies to compete in the publishing / streaming markets.
The boosters’ argument is essentially that any business that uses copyrighted works at a large enough scale shouldn’t have to pay for them, because a requirement to pay makes it harder to compete. This is clearly preposterous.
The desire to spur competition doesn’t mean we let other businesses that use copyrighted works ignore copyright. Nor should we do so with AI.
(Besides, there are ways of licensing without large upfront fees, e.g. revenue shares.)
The argument that AI training should not require licensing because it makes it harder for smaller AI companies to compete is not a serious argument. Rather, it is an attempt at a post hoc justification for what many consider theft.
This piece is just over a year old but nothing has changed , and that’s the issue. No policy or regulation changes to stop the biggest heist in global history not to mention all the fakes and “lies”. https://t.co/cLwViCcfwZ
It’s laughable that some people think liveable cities are a conspiracy to confine us !!! It’s just common sense that not having to drive everywhere is a good thing. Convenient cities become conspiracy targets https://t.co/TmjgagWfap
Meanwhile in Auckland where there are 66 business districts, the one that's just recorded the best increase in trade - 26%! - is Karangahape Rd. Where *checks notes* they built cycle lanes and took out some car parks on the street. More on this in my @nzherald column tomorrow.
Interesting thoughts and as the ODT has shown, there is nothing wrong, and its good business, to be parochial....Is consolidation really the solution to our media's problems? https://t.co/k1cw03CNrI
Very true and first time I've seen "infobesity" used in a newspaper :- David Thomason: Advertising déjà vu - Why so many Covid 19 ads feel the same, via @nzherald https://t.co/PFvWEmpv7i
And more funding for SRN would benefit all youth and community radio in NZ ......... What RNZ's 'youth network' could learn from student radio https://t.co/7pRDY7Nua7