Millions of people now use AI chatbots as a source of news to find quick answers to questions about current events.

A 2026 survey by the Reuters Institute for the Study of Journalism, which questioned more than 90,000 people worldwide, found that 10% use AI chatbots for news on a weekly basis, a figure that is significantly higher among younger audiences.

Explaining what is happening in the world, why it matters, and what it means has traditionally been the work of journalism. Media organizations send reporters into the field, commission investigations and conduct research, producing stories that help audiences understand global events. But as more people turn to AI tools for information, the relationship between those producing journalism and those distributing it is changing.

"In the past, search engines like Google indexed our websites and directed users to their articles, generating traffic in return," Julia Becker, publisher of Funke Media Group, told a panel at the Global Media Forum (GMF) in Berlin. "But now that exchange has broken down."

"AI companies draw on our quality reporting to generate answers, yet the reporting itself often remains largely invisible by the end of this process," said DW Director General Barbara Massing in a keynote address at the forum.

Their remarks reflect a worry that is growing across the media industry. Publicly available news is copied and summarized by AI chatbots, while news organizations are becoming less and less able to rely on the money and visibility that used to come from distributing that work.

Journalism and AI: A growing imbalance?

There has been growing concern about how copyrighted material is collected, used, and reproduced by artificial intelligence, and whether the companies that build those systems should pay the people and organizations that created the material.

Media houses are now pressing "Big Tech" companies to ask permission before using their work and to pay for it.

In October, German broadcasters ARD and RTL, along with the country's Federal Association of Digital Publishers and Newspaper Publishers (BDZV) and the Association of Private Media (VAUNET), called for stronger copyright protections for media companies.

In the US, lawsuits are adding pressure. The New York Times has sued internet company Brave Softwar for allegedly using its content without permission and getting around its paywall.

Meanwhile, legal institutions in the EU, including the European Court of Justice (ECJ) are examining whether AI-generated answers can themselves infringe copyright.

In February, a coalition of media organizations, including the Guardian, the BBC, and the Financial Times, called for rules that would give publishers more control over how AI systems copy and reproduce their content.

The problem goes beyond whether AI companies may use journalism. It also concerns how publishers and journalists can get paid when their work becomes part of the answers AI systems generate.

"AI systems shouldn't just use quality journalism. They should also pay for it," Massing said.

How could AI companies pay for the content they use?

Professor Martin Senftleben, who directs the Institute for Information Law at the University of Amsterdam, separates two ways AI companies use journalistic material: Training their models and generating responses to individual user questions.

Payment for training is a more complicated legal question, he said. The more urgent issue is what happens when an AI system uses journalism to answer a question in real time.

He argued that European copyright law may already offer a way to address part of the problem. Under existing EU law, publishers and journalists may be able to claim payment collectively, without waiting for a new legal framework.

"What are we waiting for? Germany is a large member state. This should be introduced in Germany," he told the forum.

One option, Senftleben suggested, is a legally required payment system in which every AI response that draws on journalism triggers a payment. A collecting organization would gather the money and distribute it among publishers and journalists, in the same way that music royalties are collected and distributed.

Once such a system exists, publishers and authors could use its rates as a starting point when negotiating individual licensing deals with AI companies.

Existing deals and concerns about too much regulation

Lutz Mache, senior manager of government affairs and public policy at Google, told the GMF panel that Google is willing to negotiate agreements with German media companies and to discuss new AI products with them.

Google has also started testing other ways to pay publishers for their contributions to AI-generated answers. Its "AI Contribution" pilot, which is being tested with selected publishers, aims to reward content that plays a significant role in answers produced by products such as Gemini and AI Overviews. But the program is still an early experiment, not a general payment system open to all publishers.

Microsoft, Anthropic, and OpenAI have recently said they support clearer, standardized signals that let website owners state whether their content may be used to train AI models.

However, not everyone thinks more regulation is the answer.

Jörg Heidrich, a lawyer who specializes in technology law and serves as AI commissioner at the German media company Heise, warned against adding new rules without considering their effects on the wider tech sector.

Europe has already fallen behind in the global AI race, he argued. More regulation, he warned, could make it harder for European companies to compete with major providers based elsewhere.

Heidrich also pointed to smaller players, including startups, independent developers, and research institutions. They depend on access to data to build and train AI systems. He argued that the debate over copyright and payment risks overlooking these groups, and that new restrictions could hurt their work.