(Credit: NurPhoto/Contributor/NurPhoto via Getty Images)
OpenAI and The Financial Times (FT) have struck a content licensing deal, giving the ChatGPT maker acess to FT articles to train its models on while providing the publication with extra cash and the potential for more site traffic.
OpenAI says the partnership is "about finding creative and productive ways for AI to empower news organizations" while enriching the ChatGPT experience. The deal has "broader implications" for the media industry, says FT chief executive John Ridding, and sets the standard that AI companies should pay publishers for their content.
FT has blocked OpenAI from crawling its articles for free since September 2023, as have most major news outlets in the US. Formal licensing deals could bring mutually beneficial collaboration, though OpenAI is reportedly offering only $1 million to $5 million, according to The Information. FT did not disclose how much OpenAI the deal is worth.
A digital subscription to FT starts at $39 per month, so the new arrangement could entice new subscribers who visit the website after teasing the content through ChatGPT. "This means that the chatbot’s 100 million users worldwide can access FT reporting through ChatGPT, while providing a route back to the original source material," says the FT.
However, ChatGPT does not link to its sources unless the user explicitly asks it to, at least on the free version. In our testing, we found that source links were more commonly included with a paid ChatGPT Plus account.
Publications have reason to be suspicious of flashy tech companies promising ever-increasing traffic and access to new audiences. Google courted publications in its early days, striking digital deals that have not sufficiently replaced revenue from print media. It's a similar story with Meta.
So the journalism industry is figuring out if and how to embrace AI. The New York Times is suing OpenAI for using its content without permission or payment, and at the same time exploring uses for it in the newsroom. Other publications, from Sports Illustrated to CNET, found themselves in hot water after failing to disclose the use of AI in its articles. (PCMag does not use AI to generate original content, per our Editorial Mission.)