As more people use OpenAI’s ChatGPT and other chatbots to find information, AI companies are trying to find ways to integrate the latest, most useful, and accurate information into their products. AI models still often contain false information. Relying on news content from third parties is therefore a way to increase the reliability of AI responses.
News organizations, for their part, fear that more and more people are using AI to get their news, which would take away traffic and subscribers from their websites and further hurt their businesses already reeling from the boom. social media.
“We believe that searches using AI models will be one of the fundamental ways to navigate the web in the future,” said Nicholas Thompson, CEO of Atlantic. announcement on the magazine’s website.
TO CATCH UP
Summarized stories to stay informed quickly
But not everyone in the publishing industry agrees. The deals represent a major divide growing within the journalism and publishing worlds, with some organizations suing OpenAI for using their content to train AI algorithms without payment, and others selling news content fresh to the AI company for a piece of the growing AI revenue pie.
Some news outlets, including the New York Times and several newspapers owned by investment fund Alden Global Capital, sued OpenAI for using their copyrighted works to train their AI algorithms, joining a group of authors, artists and musicians. Fight against what they see as a massive wave of theft from tech companies.
OpenAI and others tech companies like Google and Meta have trained the AI models that power their chatbots by scraping much of the web without payment or permission. The tech companies all say that training on scraped data is legal under fair use – a concept in copyright law that allows reuse of material if it is substantially modified.
But OpenAI also needs access to new content behind the paywalls of news sites to present its chatbots as the most up-to-date and useful – and these content deals help the company do that. Soon, when users ask ChatGPT for news updates, they’ll see headlines, article sections, and links to the news sites they’ve partnered with. the company.
Less than a week before the deal, The Atlantic published an opinion piece by Jessica Lessin, a longtime tech journalist and founder of tech news site The Information, in which Lessin argued that news organizations were naive by signing agreements with OpenAI.
“For as long as I’ve been reporting on Internet companies, I’ve watched news executives try to bend their companies to the will of Apple, of Google, of Meta,” Lessin wrote. “It never works as planned.”
