In her 1967 essay “Truth and Politics,” the political philosopher Hannah Arendt observed that “Freedom of opinion is a farce unless factual information is guaranteed and the facts themselves are not in dispute.”
The news supplied by trustworthy journalists has long been an indispensable source of that information, without which democratic societies cannot function effectively.
But as people increasingly rely on AI chatbots for news, siphoning traffic and profits from outlets often operating on a shoestring budget, good journalism is at risk of disappearing.
The paradox is that trusted news sources are essential to AI’s credibility. Failure to recognize this dependence may result in AI tools further destabilizing the already precarious journalism industry at a moment of profound upheaval and democratic erosion.
Mariya Gabriel, UNESCO’s Assistant Director-General for Communication and Information, put it well at the World News Media Congress in June: “Support for journalism is not a luxury. It is democratic infrastructure.”
Around the same time, UNESCO translated that principle into practice by unveiling its Draft Guidance on Fair Compensation for News, with the final version expected later this year.
Supporting journalism
AI firms have a responsibility to help sustain the institutions producing the information that their models train on and reproduce, because every AI-generated answer depends on knowledge that someone first had to discover, verify, report, edit, and publish.
An analysis of more than 680 million individual citations across popular chatbots such as ChatGPT, Claude, and Gemini found that journalism accounts for 27% of them, rising to 49% for queries about recent events.
Supporting journalism would help AI firms win the public’s confidence
Without verified information from outside sources, AI systems not only give inaccurate or out-of-date information, but also degrade.
Supporting journalism would also help AI firms win the public’s confidence, as just 20% of people globally trust chatbot answers. That caution looks justified.
A 2025 study by the European Broadcasting Union and the BBC evaluated over 3,000 responses from leading AI assistants and found that 45% contained at least one significant issue, and that 31% had serious sourcing failures.
Shared responsibility
UNESCO’s proposal seeks to address this problem by establishing a framework rooted in shared responsibility while protecting human rights and media freedom.
This approach, which regards journalism not simply as an industry but as a public good whose sustainability underpins access to reliable information, advocates the use of bargaining codes, competition remedies, public and platform-financed funds, and collective licensing.
These tools would be paired with transparency and attribution obligations and protections for smaller and local media.
UNESCO’s proposal shifts the debate away from copyright toward the public interest in preserving sources of accurate, trusted information
The real strength of UNESCO’s proposal is that it shifts the debate away from copyright toward the public interest in preserving sources of accurate, trusted information.
The World Association of News Publishers (WAN-IFRA, of which I am CEO) unreservedly endorses that framing of the issue.
For two decades, the internet rested on a workable bargain: publishers invested in journalism, and tech platforms helped audiences discover it.
But the arrival of AI chatbots that answer questions directly instead of sending users to the publishers who invested in finding the answers has broken that bargain.
Investment in trusted information
We must create a new value-exchange model. As the New York Times publisher A.G. Sulzberger told delegates at June’s World News Media Congress, AI companies should “support the healthy, sustainable flow of the information and ideas and creativity” that powers their technology.
This is not charity; it is enlightened self-interest. Every AI firm is one misinformation crisis away from discovering the difficulty of rebuilding broken trust.
Algorithms cannot replicate independent news publications’ rigorous verification process and professional editorial judgment.
Fair compensation should not be viewed as a regulatory burden
Fair compensation should not be viewed as a regulatory burden.
It is an investment in trusted information, democratic resilience, and the long-term credibility of AI—a technology that, contrary to the popular narrative, news publishers and journalists are embracing.
For example, investigative reporters are using AI to analyze thousands of documents and public records, helping them identify patterns, connections, and anomalies that would otherwise take weeks or months to uncover.
This gives them more time to do what the technology cannot: verify the findings, cultivate sources, ask difficult questions, and establish what actually happened.
Building a future in which both AI and journalism can thrive is a responsibility that must be shared by publishers, policymakers, and tech companies.
UNESCO has taken an important step toward making this future a reality, and its guidelines deserve broad global support.
The success of the AI ecosystem will be determined not only by the sophistication of its algorithms, but also by the continued existence of trusted, diverse, and economically sustainable news sources.
Stig Ørskov is CEO of WAN-IFRA (the World Association of News Publishers), representing 20,000 publications in more than 120 countries.
