![Dreamforce 2023 - Day 1](https://media.bizj.us/view/img/12594224/dreamforce2023day1-55*540xx3936-2214-0-205.jpg)
After freely scraping the internet's content for years to train its algorithms, OpenAI is now trying to ink licensing deals with major news and media organizations. And now we know at least a few of the publications that the company is negotiating with.
CNN, Time and Fox Corp. are among the dozens of media publishers talking to OpenAI about licensing deals, according to a new report from Bloomberg News.
Those talks include potential deals to use media content for OpenAI's flagship chatbot, ChatGPT, for both training purposes and for being featured in responses to user inquiries.
Video and other images could be included in some of the deals, in addition to text-based content, the report says.
OpenAI declined to comment specifically on the details in Bloomberg's report on Wednesday, but a representative pointed me to a Jan. 8 blog post in which OpenAI responded to a lawsuit filed by the New York Times in December over alleged copyright infringement.
In that blog post, OpenAI confirms it has "met with dozens" of news organizations as well as an industry group, the News/Media Alliance.
CNN declined to comment Wednesday.
"Our goals are to support a healthy news ecosystem, be a good partner, and create mutually beneficial opportunities," OpenAI wrote, pointing to partnerships it has already signed with the Associated Press, Axel Springer, the American Journalism Project and New York University.
The Information reported earlier this month that OpenAI was offering media organizations between $1 million to $5 million annually for licensing agreements.
However, OpenAI also maintains that using publicly available content on the internet falls into fair use.
"Training AI models using publicly available internet materials is fair use," OpenAI wrote in its Jan. 8 blog post. "That being said, legal right is less important to us than being good citizens. We have led the AI industry in providing a simple opt-out process for publishers (which the New York Times adopted in August 2023) to prevent our tools from accessing their sites."