Thursday, May 23, 2024

Open AI Being Suied For Copyright Infringement

The very substance of artificial intelligence and large language models has come under attack from The Authors Guild, an organization that exists to promote free speech, fair contracts and copyright. They are suing Open AI for using copyrighted material, ie. over 100,000 copyrighted books used to "train" Chat GPT-3. According to Digital Information World, Open AI was well aware of the illegality of what they were doing, deleted the data sets used and the researchers involved in the project are no longer available.

It's not hard to imagine that a successful suit against Open AI and other artificial intelligence efforts might throw a giant monkey wrench into the AI machinery. Certainly the opposing sides might come to an amicable and financially rewarding agreement on the matter but the possibility that negotiations could drag on for an extended period could slow the adoption of what seems to be the greatest invention since beer.

Not everyone sees AI and large language models as an unalloyed blessing. This article by Garrison Lovely points out that nobody can really balance the risk-reward situation if AI is employed as currently visualized.

Open AI update:  News Corp, owner of the Wall Street Journal and New York Post, has reached an agreement with generative artificial intelligence Open AI allowing the use of published news to answer questions and train its system. It could mean $250 million dollars for News Corp over the next 5 years and include cash and credits for the use of the Open AI technology.

It seems as though News Corp, if making use of Open AI, will be regurgitating its own work, as well as that of other publications.

 

No comments: