Delivering his opening keynote at Wan-Ifra’s World News Media Congress in Marseille, The New York Times chairman and publisher A.G. Sulzberger delivered a scathing assessment of how AI companies are “strip-mining” publishers’ sites and value, Teemu Henriksson writes.
(Full text of his speech here.)
“Some tech leaders will portray my comments today as anti-AI. As defending the old status quo. As yet another ossified institution lashing out at the innovators who are driving the forward march of progress,” he said.
Although news publishers’ “original, high-quality content is particularly valuable” for training AI models, the reality is that AI platforms’ “parasitic posture” is undercutting that value, according to Sulzberger.
AI companies – “already among the richest and most powerful in human history” – are carving out an increasingly central role in the public information ecosystem, Sulzberger said. But they are failing to “embrace a core responsibility that comes with that power – to ensure the public has access to trustworthy news and information.”
“Their hijacking of the public square is made possible by the original sin that animates their AI products – a brazen theft of intellectual property that has occurred at an unprecedented scale,“ he said.
As he mentioned, it would be easy to label him and The Times as the stereotypical disgruntled legacy brand in the midst of battling another tech disruption.
“So it’s worth stating this plainly: The news organization I lead, The New York Times, has a long record of embracing technology to advance the mission of independent journalism. We have a history of respectful partnerships with tech companies to bring that journalism to new readers in new ways.”
“Today my colleagues are using AI technology – responsibly, ethically, and with humans making the decisions – to improve how we report, edit, distribute, and monetize our journalism. Holding a powerful new technology at arms length is a recipe for failure.”
“I’m not calling AI – or the tech giants that control this technology – inherently bad or evil. I’m simply warning that AI companies are making choices, choices that violate settled law, threaten the viability of creative work, and appear likely to cause a great deal of unnecessary harm.”
Sulzberger said that AI companies are reluctant to offer compensation for their use of publishers’ journalism – which they reductively call “data”. However, they have no such reservations about paying for other elements that power their platforms.
“Talent, compute, energy and data are all essential to the success of AI and, therefore, to the success of the tech giants,” he said.
“The first three are paid for, because – of course they are. […] In contrast, AI companies take ‘data’ without consent or compensation.”
“And this theft isn’t just happening because publishers are leaving their toys out on the lawn; it’s happening when they are locked up safely in the house. A study found that about 30 percent of AI bot scrapes violate explicit restrictions on accessing and taking websites’ content, including content protected behind paywalls.”
Original journalism’s value to AI
Although publishers’ content is taken and used for free, Sulzberger said that it is actually their original, high-quality content that is “particularly valuable” for building effective AI tools.
“One of [OpenAI’s] engineers wrote that the success of models is ‘not determined by architecture, hyperparameters, or optimiser choices. It’s determined by your dataset, nothing else.’”
He compared the actions of current AI executives to the leaders of previously disruptive tech companies, who typically argued that they were building mutually beneficial relationships with content creators. For example, Spotify, “which has its critics in the music industry,” would refer to the payments it sends to musical artists, Sulzberger said.
“The AI companies, in contrast, have adopted a more overtly parasitic posture, one more akin to Napster, the old pirated music platform,“ he said. “The science fiction writer Margaret Atwood likened this dynamic to being ‘murdered by my replica.’”
The New York Times’s experience is instructive in this regard. According to Sulzberger, the newspaper was the main source of proprietary data in a dataset that was used to train various LLMs, followed by other publishers such as The Guardian and The Los Angeles Times.
“Yet the tech giants have argued consistently that they should not be expected to ask permission to use – let alone pay for – this kind of intellectual property,” Sulzberger said.
For its part, The New York Times decided to take OpenAI and its partner Microsoft to court “for brazen violations of our intellectual property rights protected by US copyright law both in the training of their models and the ongoing use of our work in their products,” he said.
“But lawsuits are slow and they’re expensive – ours has already stretched two-and-a-half years and cost over 20 million dollars. As AI companies are doubtlessly aware, most news organisations lack the resources to go to court to enforce their rights.”
Internet’s broken value exchange
While the news industry is no stranger to profound disruption caused by the tech industry – the US has lost more than 3,000 newspapers over the last 20 years – the AI disruption is poised to be even more damaging, Sulzberger said.
Previously, tech innovations came with a real – “if skewed” – value exchange: platforms offered publishers a way of reaching a larger audience in exchange for a large share of ad revenue.
But in the AI era, search engines such as Google, he said, are providing users with direct answers, with the consequence that “getting a Google user to click a link, according to industry research, is 10 times harder today than it was a decade ago,” Sulzberger said.
“Nevertheless, Google still sets the high water mark for sending readers to publishers and we can only hope that commitment will continue. Competing AI models send referral traffic at a rate that is 96 percent lower than Google search, according to one study.”
“This dynamic has, of course, led traffic to news websites to plummet,“ he said, citing Comscore research which showed that the largest newspapers they tracked had seen traffic declines of more than 45 percent over the last four years.
‘We cannot afford to be as naive this time’
When first confronted with previous digital disruptions, many news publishers – “including for a while The Times” – went along with the sentiment “information wants to be free,” Sulzberger said.
“We cannot afford to be as naive this time,” he said. “News organisations are collectively smaller and weaker than two decades ago. Tech giants are bigger and stronger – and far more willing to use their size and power.”
“Meanwhile, the AI wave itself may be bigger and faster as the technology continues to improve. Even if things are feeling fine now, remember that these early swells herald an approaching tsunami.”
“As we prepare, we must remind ourselves: information is valuable. Journalism is valuable,” Sulzberger said.
He offered concrete advice on how publishers can defend their work and build resilience against AI:
*Stand up for your rights. Publishers must actively insist their rights are respected, as “the alternative path of quietly tolerating the systematic theft of your work will eventually end your ability to continue producing it,” Sulzberger said.
*Deal carefully. While Sulzberger said it’s reasonable for news organisations to sign deals with AI companies, he urged publishers to consider if the payment actually represents fair value before agreeing.
*Push your legislators. Collectively, the industry needs to push for strong IP protections, limits to unauthorised bot scraping, and requirements for AI platforms to reveal how they use news publishers’ work.
*Join together. News publishers must work together and with other creative industries, to counteract the AI industry’s massive resources and power.
He also urged publishers to make their organizations more resilient in the AI era:
*Use AI the right way. Despite the questionable actions of leading AI companies, “there’s nothing inherently bad about AI technology,” he said. “AI can bring real value to organisations that find the right ways to embrace it, and a shift of this size will lay waste to any organisation that refuses to evolve.”
*Be a destination first. This doesn’t mean ignoring social media and other platforms where people consume content. “But to deepen those relationships – to make them loyal, habituated, and valuable – your audience must learn it’s better to engage with you directly rather than through someone else.”
*Focus on original reporting. To be a destination in an online sphere intermediated by AI, “you’ll need journalism so distinctive it has its own gravity. The heart of that is original reporting.”
*Explain why journalism matters. While AI companies have tremendous PR power, news industry must also make the case for original reporting and its importance to healthy societies, secure nations and strong democracies – “and show how the actions of the tech giants are putting them at risk.”
“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.
“We cannot allow AI cheerleaders to dominate the public conversation without interjecting to argue for the importance of ensuring a sustainable future for original journalism. We cannot watch as AI companies attempt to permanently dismantle the rights that give us control over the work we create.”
“We cannot sit by as this work is used to build replacement products that undermine our ability to earn the audience and revenue necessary to continue reporting the news.”
First published on Wan-Ifra. Teemu Henriksson (teemu.henriksson@wan-ifra.org) is Research Editor at Wan-Ifra
