gizmodo.com: The complete copyright-free-for-all approach that OpenAI took to its new AI video generation model, Sora 2, lasted all of one week. After initially requiring copyright holders to opt out of having their content appear in Sora-generated videos, CEO Sam Altman announced that the company will be moving to an “opt-in” model that will “give rightsholders more granular control over generation of characters”—and Sora obsessives are not taking it particularly well.
4 comments:
The people who are complaining make me extremely frustrated. Overall, I think it's no secret that I personally hate AI. I think all of it is bad, and I am very strict and firm about how and the methods I think it could be used ethically and in a good way as a tool. Making AI-generated videos is one of the many ways of using AI that I despise, and I’m glad at least some blocks and limits are being put in place, though I’d prefer if it just didn’t exist at all. This article just further proves how the people who make this sort of content have no creativity of their own and need to separate themselves from AI and find it. And if they don’t want to go find their own creativity, then they should create things by stealing the creativity and hard work of others. The whole situation just makes me extremely angry, and I’m at least glad there are some minor limits now. I hope the trend of limiting the technology continues.
This article just made me giggle. When I first read the headline, I was immediately overcome with joy that copyrighted characters were no longer being pulled as training material for an open AI program. There is still a long way to go until all copyrighted material is no longer being pulled for AI training, but this is an incredible step in the right direction. Then I read the article title again, and I processed that it was saying that people were freaking out about this, and after reading the article, all I can do is laugh because AI should never have been able to pull copyrighted material in the first place, so this is purely correcting an egregious error that OpenAI made. People are so stupid sometimes, and they make me angry, but in this case, they're stupidity is purely amusing. I hope to see more copyrighted materials pulled from AI training programs, and I hope that we will eventually get to a place where no copyrighted material will ever be used to train an AI.
The first line of the article that is right under the title really describes this article very well. The way that people are complaining about this obvious and completely legitimate change shows that the people who are complaining are completely in the wrong. It makes total sense that companies don't want their characters to be used for another person's profit without any of that profit going back to the people who painstakingly created said character. It really makes me want to say, "oh no you cant steal someone else's work to make yourself money, whatever will you do" to any of the people who are complaining. AI can generate art from other imagery so easily that it couldn't be easier for a person to steal content to combine together to make something quote unquote "new". The people who are complaining are only complaining because an extremely easy avenue of profit has been taken from them. The characters or images they are using to create AI art are well known that it makes it easy for people to recognize it, and use that to make easy money. I am glad that this is being stopped and anyone who is complaining is just calling themselves out as thieves and scummy people.
I fail to understand how training AI models on copyrighted materials is fair use. It is explicitly training a model to be able to recreate and use the content provided to it such that it can produce that content more accurately. With that aside out of the way, I appreciate the shift for right holders to opt in to Sora 2 as well as that there is a Sora 2 specific platform (in the sense that it is good to know that all the content you’re consuming is AI generated), but nonetheless remain concerned about what users are able to create with such platforms and wish there was heavier moderation in the prompts provided to the model for it’s creation. I realize this is a more complex issue that just simply providing certain modifiers to the program, but taking these problems seriously to mitigate harm is a concern that should always be on OpenAI’s mind.
Post a Comment