New York Times publisher A.G. Sulzberger warned that artificial intelligence companies are making decisions that could cause “a great deal of unnecessary harm” to journalism and the public’s access to reliable information, in remarks delivered at the World News Media Congress in France on Monday, June 1, 2026.
Sulzberger said companies behind generative AI systems, including OpenAI, Meta, Anthropic and Google, are failing to meet what he described as a basic responsibility tied to the data used to train their models: preserving the public’s access to trustworthy news. He argued that these companies have benefited from news reporting while also reshaping how audiences find and consume information, in part through AI-generated summaries and tools built on large language models trained on journalistic content.
He said the impact is already being felt across the news industry, which has faced years of financial pressure from falling advertising revenue and declining search traffic. Sulzberger warned that the trend could accelerate a broader decline in original reporting, with fewer journalists doing the expensive and difficult work of gathering information in the field, interviewing sources, investigating powerful institutions and providing context for major events.
The publisher said the loss of that reporting would weaken one of the foundations of a healthy society and stable democracy: the ability to produce truth, understanding and accountability through independent journalism. He urged other media organizations to speak more forcefully about the risks AI poses to the economics and sustainability of newsrooms, saying too many have been too quiet, too passive and too fragmented in responding to the technology’s rapid rise.
His comments come as publishers increasingly clash with AI companies over copyright and compensation. The New York Times was the first major news organization to sue OpenAI and Microsoft in 2023, accusing them of using its journalism without permission to train AI systems. Since then, other media organizations have launched similar legal actions against generative AI companies, including OpenAI and Perplexity.
The Times has also taken a more nuanced approach to AI inside its own operations. The company later struck a content-licensing deal with Amazon and has published internal principles guiding how its journalists should use AI tools. Its newsroom union has also made AI a central issue in ongoing contract talks this year.
Sulzberger emphasized that his criticism was not a rejection of AI altogether. He said the technology can be useful and that the Times wants to use it to improve efficiency. But he drew a line between productive use and what he sees as harmful industry behavior. Holding a powerful new technology at arm’s length, he said, would be a mistake. At the same time, he insisted that AI companies should not be allowed to weaken legal protections for creative work or replace original journalism with products that undermine the revenue needed to keep reporting news.
