AI is moving fast, and it’s no secret that it’s impacting Europe’s media landscape. Public service media embrace the opportunities AI brings to drive responsible innovation. They work every day to provide reliable media spaces through curated, trusted news offers and innovative formats.

A great example is A European Perspective, a partnership of 22 public service organisations that uses AI-powered translation and recommendation tools to make reliable news available in 13 languages. Crucially, journalists remain fully in control of editorial decisions. This project shows that AI, when used in the public interest, can strengthen Europe’s democratic conversation.

Going forward, the development of AI solutions that are built on shared European values and independent from global tech giants, will need strong and streamlined EU funding. In the media sector, this funding can be found through AgoraEU and the future EU Competitiveness Fund.

But let’s be clear: innovation alone is not enough. AI also brings serious questions that will require a thoughtful policy response.

1) How to get effective copyright protections and fair compensation?

Quality journalism takes time, expertise and money. Public service media across Europe spent 5.9 billion euros in news and current affairs programmes in 2024.

Yet generative AI systems are trained on media content, and produce outputs based on it, without fair payment or compensation. Media organisations are often unaware of these practices or powerless to oppose them in practice. This situation risks undermining the economic foundations of Europe’s media ecosystem.

Media organisations need practical, enforceable ways to opt out of AI training and retrieval-augmented generation. These opt-out mechanisms must work in reality, and they must not interfere with how content appears in search results.

Transparency is just as important. AI model providers should clearly disclose which sources are used to train and ground their models. Policymakers should also explore complementary measures to the DSM Copyright Directive, including new licensing solutions that ensure fair compensation when protected content is used.

Put simply: innovation should not come at the expense of journalism and creativity.

2) How can competition law support a pluralistic media ecosystem in the age of generative AI?

AI-driven interfaces are quickly becoming the new gateways to information.

When dominant platforms integrate AI-generated answers directly into their services, fewer users click through to original media websites. This has direct consequences for media’s revenues and visibility, and, in the long term, for media pluralism.

Existing competition tools, including the Digital Markets Act, must be used decisively to address risks such as self-preferencing or bundling dominant services with new AI features. In this perspective, the European Commission’s recent investigation into Google’s AI Overview and AI Mode is a welcome step.

This is not just about market fairness. It is about preserving a diverse and independent media landscape.

3) Why aren’t there sanctions against malicious deepfakes?

AI-generated deepfakes are already harming journalists and media brands. Identities are being misused to spread disinformation and fraudulent advertising, and women journalists are disproportionately targeted, limiting their safe participation in public debate.

Platforms erode trust by failing to act against such illegal content, while generating extensive revenues from it.

The Digital Services Act provides the framework. It must now be applied rigorously. Platforms that consistently fail to remove illegal deepfakes and malicious content should face clear and proportionate sanctions. Accountability needs to be more than a principle. It must be enforced.

4) What kind of safeguards for pluralism do we need? And how can we fight against misinformation?

AI systems increasingly shape how people access and understand news. According to Reuters Institute’s 2025 research, 7% of online news consumers use AI assistants to get their news, rising to 15% of under-25s, highlighting the rapid pace of adoption in just a few years.

However, these systems are prone to errors and can accelerate the spread of misinformation. Pan-European research by the EBU and the BBC shows that 45% of AI assistants’ outputs contain significant issues across languages and platforms, including misrepresentations of PSM news content.

AI-generated outputs must include proper attribution, clearly linking to original sources so citizens can verify information and assess credibility. Accuracy standards for news-related outputs need to be strengthened. Clear labelling of AI-generated content and the rollout of provenance standards such as C2PA can also help rebuild trust.

Most importantly, AI systems should promote diverse and trustworthy sources that meet recognised editorial standards, supporting Europe’s pluralistic information space rather than narrowing it. As AI solutions are becoming more integrated into media interfaces (e.g. connected TVs, smart speakers and infotainment dashboards in cars), Europe also needs to ensure that its regulators put in place strong regimes ensuring prominence of media with particular value for society, such as public service media, on key media interfaces.

In closing

AI has enormous potential to help Europe’s media innovate and collaborate. But without effective copyright protection, strong competition enforcement, real platform accountability and safeguards for pluralism, the same technology could weaken democratic debate.

Innovation needs investment, but it also needs guardrails.

If you found this blogpost interesting EBU Members are encouraged to sign up for our AI Forum, on 25 March in Brussels.