By Rareview
MARSEILLE, France — The publisher of The New York Times, A.G. Sulzberger, has launched one of the strongest attacks yet on the artificial intelligence industry, accusing leading AI firms of building lucrative businesses on journalism they neither created nor paid for.
Speaking at the 77th World News Media Congress in Marseille, France, Sulzberger warned that companies behind today’s most powerful AI systems are systematically using copyrighted news content without permission, threatening the economic foundations of journalism and, by extension, democratic societies.
In a speech that resonated with media executives from more than 60 countries, the New York Times publisher described the practices of major AI developers as a form of industrial-scale appropriation.
“Our profession has been too quiet, too passive and too fragmented in the face of abuses by the companies leading the AI revolution,” Sulzberger said. He argued that technology firms are “strip-mining” news websites, repackaging journalistic work, and profiting from it without compensating the publishers that funded and produced the original reporting.
His criticism was directed at some of the world’s largest AI players, including OpenAI, Microsoft, Google and Anthropic, companies that collectively command trillions of dollars in market value and are racing to dominate the fast-growing artificial intelligence sector.
At the centre of the dispute is the question of how AI systems are trained. Large language models learn patterns in language by processing vast quantities of text, including books, academic papers, websites and news articles. Publishers argue that their copyrighted material has been used without consent to train these systems, while AI companies have generally maintained that such use falls within legal doctrines such as fair use and promotes innovation.
The New York Times has emerged as one of the most prominent challengers of the industry. Sulzberger disclosed that the newspaper has already spent more than $20 million pursuing legal action against OpenAI and Microsoft over alleged copyright infringement and has also filed litigation against AI search company Perplexity. He noted that many smaller publishers lack the financial resources to wage similar legal battles.
Beyond the copyright issue, Sulzberger raised concerns about what AI could do to the business model that supports journalism.
Traditionally, search engines and social media platforms directed readers to publishers’ websites, where advertising and subscriptions generated revenue. AI-powered search tools, however, increasingly provide direct answers to users’ questions, reducing the need to visit the original source.
According to figures cited by Sulzberger, users are now significantly less likely to click through to news websites when answers are generated by AI systems, while traffic to major publishers has fallen sharply in recent years. The result, he warned, is a growing disconnect between those who create news and those who profit from its distribution.
The concern extends far beyond newspapers. Sulzberger argued that the same issues confront book publishers, musicians, filmmakers, researchers and other creators whose work may be used to train AI systems. Collectively, he said, the world’s creative industries support tens of millions of jobs and generate trillions of dollars in economic value.
“Tech giants strip-mine news websites without permission or compensation,” Sulzberger said, warning that the trend could lead to fewer journalists, less original reporting and weaker public accountability.
His remarks come amid a growing global debate over how artificial intelligence should be regulated. Governments in North America, Europe and elsewhere are grappling with questions surrounding copyright, transparency, competition and the broader societal impact of AI technologies.
For news organisations, the stakes are particularly high. Local newspapers and independent media outlets have already spent years battling declining advertising revenues and shifting audience habits. Many publishers now fear that AI could become the latest disruption to a business model that remains essential for informing citizens and scrutinising those in power.
Yet the debate is not a simple contest between technology and journalism. Artificial intelligence is already helping newsrooms automate routine tasks, analyse large datasets and improve audience engagement. Many publishers, including some of the industry’s largest players, have entered licensing agreements with AI companies rather than pursuing litigation. The challenge lies in determining whether AI and journalism can develop a sustainable partnership or whether one will undermine the other.
Sulzberger’s warning therefore speaks to a broader question facing modern societies: who pays for the production of trustworthy information in the digital age? If the institutions that finance original reporting continue to lose revenue while technology platforms accumulate greater control over information flows, the long-term consequences may extend well beyond the newsroom.
Journalism is expensive. Investigations can take months, sometimes years. Reporters travel, verify facts, challenge official narratives and expose wrongdoing. AI systems can summarise that work in seconds, but they cannot independently replace the vast infrastructure required to produce it.
As courts, regulators and technology companies wrestle with the issue, the outcome could shape not only the future of media economics but also the quality of public discourse itself. For Sulzberger and many publishers gathered in Marseille, the central concern is clear: if society weakens the institutions that generate original facts, it risks creating an information ecosystem rich in answers but increasingly poor in verified truth.