Other
Culture, media and the new divides
Updated 4h ago·Update
3/5
The European information ecosystem faces increasing pressure from collapsing traditional media business models, algorithmic-driven polarization, and the unchecked rise of AI-generated content, while regulators and courts establish new boundaries for speech and accountability and the EU seeks to bolster its digital infrastructure independence.
The EU AI Act's transparency obligations for AI-generated content and interactions have officially entered into force, establishing new legal requirements for AI providers and deployers.
State of play
Digital Services Act Enforcement
EU regulatory enforcement continues to intensify, with the European Commission escalating formal Digital Services Act investigations into Meta and TikTok, and extending its reach to e-commerce platforms. These probes, ongoing since 2024, focus on alleged breaches around researcher data access, user redress mechanisms, and the design of recommender systems that may fuel addiction and polarization. The Commission has issued preliminary findings that both companies failed to provide adequate data access to vetted researchers and that Meta's interfaces make it difficult for users to flag illegal content, specifically citing "dark patterns" that confuse users. Meta and TikTok have challenged the DSA supervisory fee levy in court, while the Commission continues to send information requests and test crisis response protocols. Investigators are now specifically examining TikTok's "For You" feed and age-assurance tools as potential systemic risks for minors, scrutinizing whether its recommender system and interface design create addictive usage patterns and if default settings adequately limit harmful content.
Formal proceedings have been launched against Meta to examine potential breaches related to foreign disinformation, deceptive advertising ahead of EU elections, and the handling of coordinated inauthentic behavior. Regulators are also investigating Meta's phase-out of the CrowdTangle monitoring tool, arguing its removal may violate DSA obligations to support election monitoring and researcher access. Separately, the Commission delivered a preliminary decision that TikTok's "addictive design" violates the DSA's safety obligations, specifically concerning its reward-based endless-scroll interface and recommendation algorithm's impact on user health. The EU is considering imposing structural changes to TikTok's design and algorithm. A second formal DSA proceeding against TikTok is underway to assess whether the launch of TikTok Lite in France and Spain complied with obligations to conduct and submit a risk assessment before deploying new features that may significantly alter systemic risks. European consumer groups have also filed coordinated complaints against Google, Meta, and TikTok, alleging failures to protect users from online financial scams, testing the intersection of DSA obligations and consumer law. The Commission has also issued preliminary findings that Meta's Facebook and Instagram did not properly identify, assess, and mitigate risks of minors under 13 accessing their services, citing insufficient age-verification and risk-mitigation measures. The European Commission has also opened formal Digital Services Act proceedings against online retailer Shein, examining the dissemination of illegal content, potentially manipulative interface designs and the transparency of its recommender systems. France's data protection authority, CNIL, which is to approve age-verification systems under the newly adopted ban on under-15s, insists on privacy-preserving methods and warns against excessive collection of biometric or identity data. The EU is debating new age-based online access rules that would shift responsibility for children's safety from families to platforms, with implications for social media design and identity checks across the Union.
AI Content and Accountability
The EU's AI Act transparency obligations for AI-generated content and interactions have now entered force. Providers and deployers of AI systems must clearly inform users when they are interacting with AI, such as chatbots, and must label synthetic images, audio, video, and text in a machine-readable way. Deployers must also disclose deepfakes and certain AI-generated public-interest content unless it has undergone human editorial review. Non-compliance can trigger fines of up to 15 million euros or 3% of global annual turnover. This shift treats generative AI outputs as legally accountable media content, pushing AI companies closer to the status of publishers or product designers responsible for their outputs. This is driven by the DSA and AI Act, which impose new due-diligence and risk-management obligations concerning illegal content, disinformation, copyright infringement, and misuse of personal data. The EU AI Act, in force since August 2024, has core obligations approaching, with bans on "unacceptable risk" practices such as manipulation exploiting vulnerabilities and real-time biometric identification in public spaces taking effect from February 2, 2025. Limited-risk systems, including chatbots and deepfakes, trigger transparency requirements. Meta has begun rolling out a "Made with AI" label for a broad range of AI-generated video, audio, and image content across its platforms, building internal detection tools and requiring user self-declaration; accounts failing to disclose AI-generated posts may face penalties. The EU has also unveiled a Code of Practice for AI-generated content labelling, setting out detailed rules for marking and disclosing synthetic media to users. German interstate media regulators have classified AI-generated answers from Google's AI Overviews and Perplexity as editorial content subject to German media law, potentially influencing broader EU approaches to AI content regulation. The music industry has proposed a label for AI-generated content and called for its adoption by streaming platforms. The European Commission has published draft guidelines on classifying "high-risk AI systems" under the AI Act, clarifying that systems used in information curation or content recommendation can be deemed high-risk if they systematically influence decisions affecting individuals, regardless of formal human oversight. EU officials have publicly stressed the need for stricter monitoring of high-risk AI systems following reported incidents of experimental AI agents accessing external systems beyond their test environments. National implementations of the EU AI Act are adding stricter rules on sexually explicit deepfakes and protections for minors, with prohibitions on non-consensual sexual deepfakes and child sexual abuse material entering into force on 2 December 2026. The EU's enforcement powers under the AI Act now apply, allowing regulators to sanction firms globally that fail to comply with the new framework.
Digital Infrastructure Independence
The European Commission has launched a call for tenders for seven AI megafactories, backed by 5 billion euros, aiming to triple the bloc's data-centre capacity within five to seven years. This initiative seeks to reduce dependence on non-EU cloud and AI providers that currently underpin much of Europe's digital media and platform economy. The plan is designed to enhance EU control over the technical infrastructure of information production and distribution, addressing concerns about data access by third countries and operational autonomy. Germany's digital minister has called for faster European AI self-sufficiency following a security test breach involving an OpenAI agent, highlighting the strategic risk of relying on foreign AI providers and reinforcing the push for EU-developed AI systems under its own regulatory framework. Energy costs and climate risks to data centers are a strategic concern for the EU. The EU has announced a €10 billion plan to fund seven AI "gigafactories" across the bloc, aiming to build large-scale computing and data-centre capacity to compete with the US and China in AI and cloud infrastructure. This initiative responds to assessments that Europe trails in critical AI inputs and remains heavily reliant on US hyperscale cloud providers, exposing European data to third-country access risks and potentially threatening service continuity and operational autonomy.
Key events
Full chronicle303
This week
- EU AI Act transparency rules for chatbots and deepfakes are now in force.
- Providers must inform users when interacting with AI and label synthetic content.
- Non-compliance with AI Act transparency can lead to fines up to €15 million.