What are the ethical implications of using articles, images, and music, found online, to train AI?
Moral dilemma question of the day ?
Since AI is being trained by articles, images and art, and music, found online, what are the implications of using those works without paying the creators? Are there not risks in using unauthorized or illegal works?
Much like how a person can cite artistic works, I feel AI learning should have similar boundaries. Considering that the AI itself is a piece of IP designed by a team that could be considered artists themselves, the intent of said work has major implications.
If the AI is training on other’s IP for the purpose of protection and detection of misuse, by all means, give it all the data.
If the intent is for the AI to learn how to imitate and create it’s own variation of the art, that is morally inconsiderate without compensating the original artist(s).
I just google that, and found one question in Quora, ‘Is it ethical to pirate educational media?’. One answer popped up ‘Yes. All education should be free.’, although both questions and answers are updated like 10 years ago.
However, if we take AI training in education viewpoint, there would be another question, is it moral to educate AI by misinformation?
Back to Paul’s question, the AI training process is similar to the human browsing behavior. Thus, is it legal to let AI browse unauthorized or illegal works?
