Media vs Big Tech: Battle for AI survival flares up in Germany
European media demand control over their data in the AI era. “Without intervention, a scorched-earth collapse in journalism threatens.”
Published on April 26, 2026
Team IO+ selects and features the most important news stories on innovation and technology, carefully curated by our editors.
The rise of generative artificial intelligence is placing a heavy burden on the European media landscape. Major technology companies use millions of news articles every day to train their AI models. They generally pay nothing for this. These systems then present the collected information to the user in ready-made summaries. This leads to a direct decline in the number of visitors to the original news platforms. Five major German media organizations are now sounding the alarm. They demand strict rules to protect journalism against the market power of global AI giants. Without swift political action, irreparable damage threatens Europe’s information supply, these German media say.
The core of the problem lies in the way search engines are changing. Traditional search engines direct users to news websites via links. New AI-driven search interfaces give direct answers. A recent report predicts a 43 percent decline in web traffic to news websites due to these AI summaries. This strikes at the heart of publishers' revenue model. Technology companies are building their AI services on the foundations of expensive journalistic productions, according to the German open letter. They themselves do not invest in editorial teams or fact-finding. The German media alliance, including ARD and ZDF, states that this leads to an unacceptable shift of attention and value.
Journalistic media risk being reduced to anonymous data suppliers for major tech platforms. This endangers democratic opinion formation. Quality journalism, after all, requires considerable financial resources. If revenues dry up, independent news gathering disappears. The European Commission takes this threat seriously. At the end of 2025, it launched a competition investigation into Google’s AI overviews. This investigation must determine whether the company is abusing its dominant market position.
Copyright and the opt-out illusion
Publishers now demand full control back over their publications. They want to decide for themselves whether AI companies may use their texts for training algorithms. Major technology companies often reject this demand. In the United States, they rely on the “fair use” doctrine. They argue that training an AI model is a transformative process. In Europe, they use the exception for “Text and Data Mining” (TDM) from the Copyright Directive. This law permits the scraping of public data, unless the creator explicitly forbids it. Publishers are now trying to enforce this prohibition technically. For this, they use specific metadata fields of the International Press Telecommunications Council (IPTC). In this way, they indicate in a machine-readable manner that their content may not be used for AI training. In addition, the C2PA standard is rapidly gaining ground. This standard adds a cryptographic watermark to files. This watermark records origin and edits beyond dispute. Refusing permission, however, is only the first step in a complex legal battle.
The battle for fair compensation
When AI companies are allowed to use journalistic data, the media houses demand fair compensation. The European Parliament supports this demand. In March 2026, parliament adopted a resolution on copyright and generative AI. This resolution rejects the idea of a one-time buyout sum. Parliament instead argues for immediate and proportional compensation. Publishers must retain the freedom to determine their own rates through voluntary licensing agreements.
The resolution contains another crucial element for the sector. There must be a mechanism for retroactive compensation. AI models have, after all, already processed immeasurable amounts of unpaid data in recent years. The German broadcasters and publishers are calling on their government to quickly anchor these European principles in national legislation. They see this as the only way to stop revenue leakage from the journalistic ecosystem. Without an enforceable compensation model, the financial incentive to produce high-quality news reports disappears. A fair compensation model restores balance in the digital economy. It forces AI developers to internalize the real costs of their data sources.
Visibility in an AI-dominated web
Besides compensation, discoverability is an existential condition for news media. AI gatekeepers are increasingly determining which information reaches citizens. The media alliance, therefore, demands competition-law rules to guarantee the “prominence” of original sources. Users must be able to see immediately where the information comes from. Experts are currently developing transparency labels for AI content. Measurement instruments such as the Source Integrity Score are central here. The Attribution Visibility Index also plays an important role. This index measures whether the original source is clearly visible to the user.
The enforcement of this visibility is governed by European competition law. Supervisors monitor, for example, the click-through rate to the original source. They require direct links on the first screen of an AI answer. In addition, experts propose giving reliable sources priority in AI summaries. For this, they want to use the Journalism Trust Initiative (JTI) indicators as an objective benchmark. These technical and legal frameworks must prevent AI systems from definitively cutting the bond between journalist and reader. This guarantees a pluralistic news supply in which citizens retain access to diverse perspectives.
Legal clashes and the AI Act
The call for strict rules clashes with the slow pace of political reality. The European Union has indeed taken a pioneering role with the AI Act, but implementation is proceeding sluggishly. The law imposes heavy fines for violating transparency rules. These fines can rise to 7.5 million euros or 1.5 percent of worldwide turnover. Article 50 of the AI Act, which regulates transparency, becomes enforceable in August 2026. At the same time, the proposed Digital Omnibus package is causing a delay. This package wants to postpone the implementation of certain obligations for high-risk AI until the end of 2027.
Germany does not want to wait for this. The German government aims to implement faster and stricter transparency requirements through an update to the so-called Digitaler Medienstaatsvertrag. This national acceleration, however, creates a legal conflict. It may clash with the European “country of origin” principle. Member states may not impose extra requirements on services that already fall under the European AI Act. This legal vacuum creates major uncertainty for both publishers and technology companies. A fragmented approach, moreover, weakens Europe’s negotiating position vis-à-vis the global tech giants.
European autonomy
The outcome of this conflict determines more than just publishers' profit margins. It touches the core of European strategic autonomy. If American or Asian AI models dominate the European news supply, the continent loses control over its own information flows. Dependence on a few large gatekeepers makes society vulnerable to disinformation and manipulation. A robust framework is therefore essential. Copyright law, competition law, and media law must complement one another in this regard.
The mandatory application of standards such as C2PA offers a technical basis for transparency. This must go hand in hand with enforceable collective licenses for data use. The coming months are crucial. The European Commission and national governments must close the loopholes in legislation. They must restore the balance between technological innovation and the preservation of an independent press. Only with a fair playing field can Europe build a digital future in which quality journalism survives, and the citizen remains assured of reliable, verifiable information.