New York Times arrives in new tech with AI lawsuit

Good news. Some press organizations are going against the journalist’s creed: never become the news. The New York Times is suing OpenAI and its biggest backer Microsoft, claiming the firms used copyrighted articles without permission as training fodder for the models that power artificial intelligence chatbots like ChatGPT. No matter the legal outcome, the case will help battered newsrooms regain some power in their relationship with big tech players.

With this suit, the Times joins the ranks of influential authors and artists throughout the U.S. seeking legal recourse from AI model-makers that they say illegally siphoned intellectual property. High-quality data fuels AI models like OpenAI’s ChatGPT, which uses the information to cobble together humanlike responses to user queries. According to the Times’ complaint, filed on Dec. 27, months of negotiations with OpenAI failed to produce an agreement that would have let them license content for a fee. The Times case goes further than its predecessors, citing several examples of ChatGPT spewing responses that appear almost identical to its articles and inaccurately attributing information to the outlet.

The Times and others have every reason to go after big pockets. Traditional news rags are having a hard time finding their footing. Analysts expect the company controlled by the Ochs-Sulzberger family to report 2023 revenue that is flat compared to last year, according to estimates from LSEG. Forays into non-news ventures, like the purchase of game Wordle, haven’t yet produced meaningful growth. And advertising is under threat as social media companies like TikTok continue to grow and take share of eyeballs. Plus firms that had helped push new readers, like Elon Musk’s X, are testing the relationship they have with news outlets.

Still lawsuits like this one might make companies like OpenAI more cautious about what they take from the internet without paying. Musk has similarly gone after firms for scraping data from X. And for the Times, it could even be the impetus for a decent deal. Proposed legislation in Australia in 2021 forced Google and Meta Platforms to agree to share some of their advertising revenue with news publishers. Rigorously fact-checked news articles are even more crucial to artificial intelligence agents, which derive value from accuracy. Bloomberg reported last month that Business Insider publisher Axel Springer inked a deal with OpenAI to make “tens of millions” of euros over three years for its data. With Times’ revenue expected to clock in at just $2.4 billion this year, every little bit counts.