Fact Checked by 
Vox Media and The Atlantic have become the latest in a long line of media organizations to sign a partnership deal with OpenAI, allowing their content to train its artificial intelligence (AI) models.
According to a press release, Vox Media will use ChatGPT to enhance its affiliate commerce product, The Strategist Gift Scout, and OpenAI will gain access to Vox’s entire archive, to enhance its training data. The Atlantic has stated that they will be developing an experimental microsite to determine how best to incorporate ChatGPT into their process while ChatGPT gains access to their archive of content.
OpenAI versus copyright lawsuits
In a bid to protect itself from copyright lawsuits and expand the training data available to its flagship large language model (LLM) ChatGPT, OpenAI has sought out and signed deals with several large media groups over the last few months.
Axel Springer (owners of Business Insider and Politico, among others) signed a deal allowing OpenAI access to summaries of its articles so the chatbot has a better chance of dealing with queries about current events. News Corp, the Murdoch news conglomerate, has signed a deal worth a reported $250 million over five years.
OpenAI has also been seeking out platforms with huge quantities of user-generated content to partner with. They currently have an API and content deal with Stack Overflow, a site where software developers exchange advice and answer each other’s questions, and a deal with Reddit worth an estimated $60 million per year (though Google’s partnership with Reddit to ‘enhance its search results’ has not been a success).
These partnerships are an attempt by OpenAI to protect itself against lawsuits for breaching copyright by using content on the web as training data. Given the fact that they are currently facing lawsuits by the New York Times, and by a group of authors and artists, it is in OpenAI’s interests to secure deals rather than contend with further legal concerns, while continuing to expand its corpus of training data.
Featured image: generated by Ideogram
