(15 Jul 2024) There are lawsuits for every type of creator. Journalists, novelists, visual artists, musicians, programmers, and actors all have lawsuits in their name. However, all the lawsuits ask one fundamental question: When, if ever, is it acceptable to train an AI on copyright-protected work without permission?
This seems like an existential threat to AI, and in many ways, it is. Sam Altman, the head of OpenAI, famously said that creating useful AI models is “impossible” without copyrighted material.
However, that is not entirely true. As we discussed last year, both Adobe and Nvidia have launched AI systems that are trained on licensed content. Though Adobe faces controversy for underpaying for its work and seeming rights grabs in its terms of service, it proves that such a system is theoretically possible.
As concern grows over AI, many are looking to use copyright to limit or even stop its growth. However, that is not going to work. Though a copyright defeat would significantly impact AI, it wouldn’t kill it.
Find out more here.
