As AI companies shift toward quality inputs, news publishers gain rare leverage to shape sustainable content licensing models.
The rapid evolution of artificial intelligence is forcing a long-overdue reckoning in the news media industry. As AI content licensing moves from an afterthought to a central pillar of the information economy, publishers are being urged to rethink how their work is valued, protected, and monetised.
That warning came during a recent webinar hosted by the INMA Product & Tech Initiative, where Dr. Jonathan Roberts, chief innovation officer at People Inc., said the future of AI depends on fixing how information is paid for.
“Solving the information economy — how we get paid and how we pay our writers — is necessary for there to be an AI future,” Roberts told publishers. “This is a problem for all of us.”
A turning point for AI and information
Roberts said the shift from 2024 to 2025 marked a fundamental change in how AI companies view information. For years, Silicon Valley operated under the assumption that training large language models on the entire internet would remove the need for new, high-quality content.
That assumption collapsed as accuracy issues and hallucinations persisted.
“We now know that the best inputs lead to the best outputs,” Roberts said. “The future of AI requires quality information, not just mass quantities of data.”
This realisation has repositioned publishers as essential partners in the AI ecosystem. Much of the internet, Roberts noted, is not factual — which creates a growing need for authoritative, nuanced, and well-sourced journalism.
“In the future, we’re going to need more content, not less,” he said.
What changed in 2025
Two major developments accelerated this shift. First was the emergence of the Chinese AI model DeepSeek, which matched leading US models at a fraction of the cost. Within days, the open-source community replicated it, proving that massive capital investment alone cannot sustain competitive advantage.
The second shift was the widespread adoption of retrieval-augmented generation by major AI players. Companies including OpenAI and Meta began grounding AI responses in cited sources to combat hallucinations.
“If you don’t cite your sources, the answer is untrustworthy,” Roberts said, noting that this principle has long been obvious to journalists but not to technology companies.
Together, these changes placed high-quality, rights-cleared content at the centre of AI performance, giving publishers leverage they have not had in years.
Why AI cannot function without publishers
Roberts broke down AI systems into three core inputs: chips, power, and information. Chips and power are commodities. Information is not.
“If you only have chips and power, you have a data centre with nothing going through it,” he said. “The thing that goes through it is words.”
Yet while chips and energy attract massive investment, information remains undervalued. Open-source models with access to the web already outperform closed models without it, underscoring the value of high-quality content.
Despite this, many AI companies continue to scrape publisher material without compensation. Roberts rejected the idea that platforms cannot afford to pay, arguing that they already monetise content indirectly.
AI developers routinely charge between $10 and $50 per thousand searches that require web access — a range comparable to what publishers earn from human traffic.
“They’re already making money from your content,” he said. “They’re just not paying you.”
Building a real market for AI content
The next challenge is creating an AI content market that works for both sides. Roberts said publishers must be paid fairly, but AI companies must also see clear benefits.
The solution, he argued, is not a blunt tax but a structured marketplace that rewards quality over quantity. How that market is designed will directly shape what content gets produced.
“Leaving it to chance is naive,” he warned.
Roberts outlined several emerging models, including paid access tollgates, revenue-share systems similar to Apple News, collective rights organisations modelled on ASCAP or BMI, and hybrid licensing frameworks. No single model will dominate, but all must be built on clear rights, transparent pricing, and fast retrieval.
Defining the next information economy
Looking ahead, Roberts said the AI era will require richer metadata, clearer authorship, and stronger signals of authority. AI systems need to know when content was created, who wrote it, and whether it is trustworthy.
Publishers, he argued, must help define those standards rather than leaving them to technology platforms.
He pointed to early initiatives at the IAB Tech Lab, Microsoft’s content marketplace, and the RSL collective as foundations for a rights-based ecosystem.
“This is a unique moment to get in the game and determine what the answer is,” Roberts said.
A rare moment of leverage
For an industry battered by years of platform dominance, the rise of AI content licensing represents both a risk and an opportunity. AI cannot deliver reliable results without high-quality journalism, and that dependency gives publishers real negotiating power.
Whether the news media industry can convert that leverage into sustainable business models may determine not only its own future, but the credibility of AI itself.