?

Storyteller's den

StorytellerStoryteller

Information about Literary Universe development and other random things that I can consider worthy of writing about.

Can AI generated works be copyrighted?

Written by Storyteller
3 years ago
0 0

With all the talk about AI and its increasing capabilities one big question came to mind. Can the works generated by AI be even copyrighted?

The obvious answer would be that it is copyrighted by the company that has created the software or AI, unless the entity putting in the prompt has a different agreement. But the problem comes when you realize that in order to create/learn the AI data from all across the Internet and human history has been used. Most of which is copyrighted. This data has been used without any compensation to the copyright owners. Honestly that is not even feasible in any practicality as pretty much everyone using the Internet would have to be compensated.

This leads me believe that OpenAI, ChatGTP and similar should be open sourced as a compensation to the copyright holders at the very least.

But it also brings an interesting question for works generated by these programs. Although the text might be novel and generated upon an original input, it could generate the work only due to having access to the sum of human knowledge on the Internet. Many noble things will be done with this technology and we will see some amazing things and stories with it, even great cooperation between AI and human to create something truly amazing. Yet I do not believe that they should be allowed the same monetization and protections as work generated by humans.

First of, the obvious, I believe in compensating those who put the time and effort to hone their craft. Using AI might be fun and will most likely be the trend for a while, but we must make sure that we still have storytellers left at the end of this period when the chaos of new technology passes over.

Going of the first point, to the second is that we need the chaos of the creation. As AI is right now, it shows us the mirror of gathering potentially the best of humanity to generate new stuff by referencing all that came before. Maybe it might not seem like it now, but I believe that in due time it will get there (and it won't be long). But after some time the prompts will have to become more complex to generate new stuff without repeating a lot of the previously generated work or it will spiral into ineligible nonsense. The reason being as there will be lack of new materials to work with from humans so it will spiral into repeating itself or get into spiral of consuming its own text.

Hence I believe that AI generated content should be accessible for free and/or under open licenses or any profit made out of it should be should, at least in part, go to support human authors.

AI is here and it will only get better. Fighting change makes no sense. It will touch many fields and we should start talking how to manage its impacts and start preparing (I'm writing this in the middle of adding an indicator to stories and universes here at Literary Universe to indicate if a work has been generated or partially generated by AI). At least for the publishing industry I think we can take some relatively easy steps to prevent the worse outcome. Take this as my opening thought in this debate.

AIcopyright

Comments