AI firms are paying millions for journalism — so why are many reporters still skint?

AI companies are striking multi-million-dollar licensing deals with some of the world’s biggest publishers, but as newsrooms scramble to monetise decades of reporting, many experts fear the AI boom could leave smaller publications and freelancers struggling to survive, write Michael Leidig and Rob Hyde

Artificial intelligence (AI) companies are suddenly paying staggering sums for journalism after years of scraping news content for free.

In an apparent gold rush, the last couple of years have seen the industry buzzing with blockbuster agreements ranging from Google’s reported US$60 million-a-year deal with Reddit to OpenAI’s five-year News Corp agreement said to be worth more than US$250 million.

Meta, Amazon and other giants have also signed partnerships with publishers including the New York Times, Financial Times, Le Monde, Associated Press and Axel Springer.

At first glance, the message seems simple: AI companies need journalism and are finally willing to cough up for it.

But despite these eye-watering deals between the giants of AI and news publishing, can ordinary journalists and smaller publishers realistically make a fair buck too?

Some legal experts suggest there are grounds for hope when it comes to journalists and small publishers standing up to AI giants, particularly under EU law.

Eleonora Rosati, Professor of Intellectual Property Law at Stockholm University, specialises in AI and copyright law.

Speaking to The European, she said: “There is definitely a market for this type of content, given its quality and timeliness, which are both key to high-quality training results.”

And she pointed to growing legal and political pressure across Europe, which should work in favour of those producing content.

“Regarding individual journalists, they would be entitled to part of the remuneration generated by press publishers when negotiating deals pursuant to their press publishers’ right under Art 15 of EU Directive 2019/790.”

But others are far more sceptical.

Ulrike Langer, an AI in journalism analyst, has spent years covering media innovation, AI and publishing strategy for German-speaking media executives and publishers.

She said the current licensing market already has a clear hierarchy.

“The market has two tiers. The top tier is real: Reuters, AP, AFP, and the Meta-News Corp deal involve serious money for structured news feeds.

“The second tier — everything below the global agencies and the largest publishers — is mostly still a conference talking point.

“Industry-wide deals may cover thousands of outlets, but there is little evidence they deliver meaningful revenue to smaller publishers.

“The market is genuine where it exists. But it does not yet exist for most of the industry.”

But Langer also argued that AI firms may ultimately value types of journalism many publishers currently overlook.

“AI companies want what they cannot already get from the open web: underrepresented places, non-idealised contexts, court records, council minutes, regional language.

“That is a structural advantage for local and specialist newsrooms over the major brands, if they have done the work to make their archive licensable in the first place.”

Right now, small and regional publishers are reopening old databases formerly treated as little more than digital storage as they try to see if decades of reporting are now valuable AI assets.

But how likely is it that their content is enough to tempt the titans of the AI industry to open their wallets?

Casey Newton, founder of influential tech newsletter Platformer, warned that the economics still overwhelmingly favour companies with enormous quantities of material.

“My impression is that there is a market both for archival content and for real-time content. But archival content doesn’t pay as well.

“The reason is that Large Language Models (LLMs) are now so large that even a relatively large collection of archival material will still make up less than one per cent of the training data of any model.”

And James Grimmelmann, a law professor at Cornell Law School and Cornell Tech, and one of the leading academic experts on generative AI, copyright law, platform regulation and digital media economics, provided an equally sober outlook.

“There is not an individual market for licensing content to AI companies. The datasets they use are so large that any individual’s content could be removed without affecting the dataset’s utility.

“AI companies will simply remove the content rather than negotiate over the details. Only large media entities have the scale of content available to make negotiation and compensation worthwhile.”

Mark Lemley, William H. Neukom Professor at Stanford Law School and director of the Stanford Program in Law, Science, and Technology, said the current licensing market for model training remains “largely limited to either high-profile news sources like the New York Times or to entities like Getty Images that can aggregate large amounts of content.”

But he believes newer AI systems based on retrieval-augmented generation — known as RAG — could prove more interesting for publishers.

“Companies using RAG may need to license content from all news sources. That could force AI firms into ongoing relationships with publishers rather than one-off scraping.”

Some smaller publishers believe specialised journalism could ultimately prove the way forward.

Isabelle Szczepanski is co-founder and editorial director of Paris-based media and technology publication ElectronLibre, which is run by just three journalists and a tech specialist.

She believes publishers may ultimately find more opportunity in AI systems built around continuous access to specialised journalism rather than one-off training deals.

“At ElectronLibre, we have taken a different approach by developing our own AI system based on retrieval-augmented generation.

“The system allows subscribers to ask questions and receive answers grounded in the publication’s archive of reporting.

“Licensing models that provide continuous access — not only to archives but also to up-to-date content — seem both more practical and more valuable for end users.

“Content that is niche, analytical, or based on original reporting likely has greater value than widely duplicated general news. Structured content significantly enhances this value.”

And others agree that the long-term winners in AI licensing may not necessarily be the largest publishers but the most trusted and specialised.

Petra Rulsch, a Dubai-based media strategist focused on AI and technology communications, said: “In the long run, I suspect the decisive factor will not simply be scale, but specialisation, credibility and structure.

“Journalism may ultimately become part of the trust infrastructure underpinning future AI ecosystems.”

But others warn that AI will simply recreate the same concentration of power that obliterated much of the news industry during the platform era, when Google, Facebook, YouTube and later TikTok became the dominant gateways to online news and advertising.

This warning resonates deeply with many journalists who watched in horror as the social-media revolution devastated advertising revenues across local and regional newsrooms.

A 2024 study by the CREATe Centre found that 93 per cent of freelance journalists had never received income from platform licensing deals.

For many freelancers, the fear is not only that years of reporting may already have been absorbed into AI systems without compensation, but that AI-generated summaries and answer engines could reduce direct traffic to news websites in much the same way social-media platforms weakened publishers’ direct relationships with readers.

Yet blocking AI systems entirely may carry long-term risks to publishers, making them invisible to the AI systems that are increasingly shaping how people discover information online.

Jeff Jarvis, professor at the Craig Newmark Graduate School of Journalism at the City University of New York, said: “If AI models are not made aware of a publisher or author, then it will not know to include that source in its answers.”

For some journalists, however, the issue goes beyond economics.

Ulrich Hottelet, a German freelance journalist specialising in artificial intelligence, IT security and data protection, has worked with organisations including Siemens, IBM and the German Federal Ministry for Economic Affairs.

He said journalists’ work must be respected in this debate.

“Without our intellectual and creative work, the billions in revenues generated by OpenAI, DeepSeek, Anthropic and others would not be possible.”

But Wendalyn Nichols, a publishing strategist and AI-content licensing specialist who previously led online Cambridge Dictionary content operations at Cambridge University Press, argued that many publishers may already have misunderstood the nature of the market.

“For the LLM builders, I would argue that the licensing opportunity is already gone.

“Once an organization has trained its AI algorithm on your data, it doesn’t need your data any longer.

“So there is just no reason to renew a licence….the algorithm cannot unlearn what it now has learned.”

Michael Leidig is a British journalist based in Austria. He was the editor of Austria Today, and the founder or cofounder of Central European News (CEN), Journalism Without Borders, the media regulator QC, and the freelance journalism initiative the Fourth Estate Alliance respectively. He is the vice chairman for the National Association of Press Agencies and the owner of NewsX. Mike also provided a series of investigations that won the Paul Foot Award in 2006.

READ MORE: ‘Why Britain still needs reporters in the courtroom‘. Britain’s tradition of open justice depends on reporters sitting in courtrooms and witnessing what happens. But as court reporting declines, a vital democratic safeguard is quietly disappearing, writes Michael Leidig.

Do you have news to share or expertise to contribute? The European welcomes insights from business leaders and sector specialists. Get in touch with our editorial team to find out more.

_{Main Image: Beyzanur K./Pexels}