Who owns Generative AI output ?

Who owns the Generative AI output ? Is it the AI company that owns the AI system or the user who gave the triggering words that triggered the output or the public whose data is being manipulated by the AI companies that have collected this data by hook or crook ?

This is a big question that needs to be addressed urgently in the near future. The AI algorithm uses the data of the public, conditioned by the textual prompt provided by the user to generate the output. The generative AI image on the right is taken from the app Dream AI.

 

I asked this question to many users and my students, they were equally unclear as to who rightly owns this.

 

The data was taken from the users, not from one user, but from millions of users across the world over the past thirty years and it will be really difficult to pinpoint which customer gave what data exactly to decide on the financial reparation to them.

Appel et al in their HBR article, Generative AI has an intellectual property problem (click here) has rightly pointed out this issue. 

As per the article and legal interpretation, the term fair use is very confusing. 

fair use doctrine, (which) allows copyrighted work to be used without the owner’s permission “for purposes such as criticism (including satire), comment, news reporting, teaching (including multiple copies for classroom use), scholarship, or research,” and for a transformative use of the copyrighted material in a manner for which it was not intended.

  To make it fair for both the customer and the artist or copyright owner, the copyright owner should be asked before the data is used for the process of generative AI. Apps like Stabble Diffusion, Midjourney have taken the dataset called LAION 5B data set which was assembled with millions of points of data got by scraping the web. Is it legal and ethical to do so is a serious question all of us need to be worried. While AI companies will take necessary precaution in their future versions of the generative AI app, like Stable Diffusion has made it clear, the content owners also need to be clear of what items of work will be out in public domain and which will not be. 

Businesses need to mention of this protection from legal complications while purchasing generative AI output from providers that the companies are bound by the laws of the land and is sure that the owners rights regarding use of copyrighted material is properly and legally taken care of.

The only solution to this seems to be generative AI generated creative art sourcing images from open source images available across the net.

Will the canvas expand to include more legalities and rights in the coming years only time can tell. 

George.