by the CLI Team
For those of us in the United States and all those ex-pats, today we celebrate our Thanksgiving Day. Happy Thanksgiving!
To everyone, we hope you enjoy the day. It is a great opportunity to reflect on our lives and what we have. We are thankful to you who receive our newsletter and read the articles. We hope you find it helpful, and we welcome your feedback on how we can make it better for you.
As this is Issue 5 of the Content Licensing Brief, and Thanksgiving for most of our readers, we thought we would bring you articles and other content from around the web discussing content licensing and artificial intelligence. Future issues of the Brief will explore licensing as a brand builder and a business development opportunity. We have lined up leaders in content licensing for one-on-one interviews and other articles. Enjoy.
How much do Google and Meta owe publishers? Twelve billion dollars, a new study says. – Columbia Journalism Review (cjr.org), Mathew Ingram, Columbia Journalism Review
This article asks a great question, “what is the value of news content to the major platforms?” Whether news, magazine or other content, what is the value to major platforms such as Google, Meta or any generative artificial intelligence is the big and tricky question. Note the Columbia University study.
The Clash Of Generative AI And Intellectual Property Law: What It Means For Businesses, Kyle Wiggers, Forbes
There is litigation going on as we write this article that is testing the use by AI and language learning models use of copyrighted materials and the rights of the copyright owner. Does this usage constitutes copyright infringement? AI will argue that it is Fair Use. Fair Use is a defense against infringement lawsuits, it is not a permission to use copyrighted materials with out a content license. Even the U.S. Copyright Office is struggling with this issue.
AI’s intellectual property battleground: word embeddings (datawellness.io), Jay Krall, Data Wellness
In this article, Jay Krall asks a great question, “what constitutes language and intellectual property.” Here Jay injects a computational linguistic technique of “word embeddings.” He gets into concepts of “word vectors” and “training documents” as a creation of AI and the property of the copyright holder, but, as he points out, copyright holders disagree. Mr. Krall asks another great question, “can a statistical representation of an authors voice imitated for commercial purposes without infringing on copyright?”
The News/Media Alliance white paper on generative AI and the infringing of copyright. This is an important report in the understanding of the challenges AI is creating to the protection of intellectual property. A must read.
AI Companies Are Running Out of Training Data (futurism.com), Maggie Harrison, Futurism
And then, what happens when AI runs out of publishers’ data? Closer to home; what happens to publishers when AI has consumed all of their data?
