“AI is bringing remarkable innovation and many benefits for people and businesses across Europe, but this progress cannot come at the expense of the principles at the heart of our societies,” said EU competition commissioner Teresa Ribera. That sentiment now frames the intensifying clash between Google’s rapid AI search expansion and Europe’s tightening regulatory grip.
Google is pushing a significant overhaul to its AI Mode, the conversational search layer that blends generative answers with traditional results. The update will embed more in‑line links and short AI‑generated “contextual introductions” explaining why a source is relevant, positioned above a carousel of cited articles. Additional linked phrases will be woven into AI responses to create clearer navigation paths to original material. More than a UI tweak, this reflects mounting pressure from Brussels to make AI‑powered search more transparent and accountable in how it surfaces and attributes publisher content.
The probe centres on whether Google’s AI Overviews and AI Mode use the works of web publishers and YouTube creators without “appropriate compensation” and without a real opt‑out. Current mechanics force publishers to choose between being excluded from AI features or losing visibility in traditional search. Technical standards like Really Simple Licensing 1.0 aim to solve that, allowing granular control so content can remain in standard search while being blocked from AI training or generative use. Backed by more than 1,500 media organizations, RSL extends robots.txt into a licensing layer, supported by infrastructure providers like Cloudflare and Akamai. Yet Google has not implemented such selective opt‑outs, a gap regulators are now probing.
Meanwhile, Google is testing the AI-powered tools with top publishers such as The Guardian, The Washington Post, Der Spiegel, and El País. In one experiment, AI-written article overviews go directly onto the participating outlets’ Google News pages; in another, real-time updates from those publishers are piped into the Gemini app. Those deals come with direct payments to offset any traffic decline-a crucial concession given the evidence of, say, the Daily Mail claiming a 50% drop in clicks since AI Overviews launched. Going global is the new “Preferred Sources” feature, which allows users to give favored outlets top billing in Top Stories, while subscription content is surfaced in its own carousel. Technically a simple thing, it’s personalization that risks creating ideological silos-a prospect media analysts have already sounded the alarm on.
Under the hood, AI Mode’s architecture combines large language models with retrieval‑augmented generation, or RAG, pulling in live web data to ground responses. The cited sources and contextual introductions will hopefully prevent hallucinations and make the responses more trustworthy, but they also expose the provenance of both the training and grounding data-that is, precisely those aspects where the EU’s DMA and copyright policy meet. The recent EU Parliament report on compulsory copyright licensing of AI training concludes that voluntary opt‑in markets are afflicted with high clearance costs and fragmented datasets, whereas opt‑out regimes reduce access without restoring incentives for new content. Its suggested statutory license with modest royalties would yield $14 billion more annual welfare than alternatives but would deprive creators of any right to refuse use entirely.
Google’s regulatory exposure does not stop at AI. The Commission is also considering a significant fine in early 2026 over compliance by Google Play with DMA rules. Investigators have pointed out technical limitations imposed on app developers to prevent them from diverting users to cheaper offers outside Play and service fees that they regard as excessive for the acquisition of the first customer. Tweaks made in August to fee models and linking rules are still inadequate, with Apple’s June App Store revamp seen as the benchmark for compliance. DMA infringements can be fined up to 10% of global annual revenue, so this is a multibillion‑euro potential liability.
Compounding matters, another suspected breach of the DMA is a separate case that accuses Google of placing its own vertical services-Shopping, Hotels, and Flights-at the top of search results. Google has offered a set of incremental changes, but the Commission claims these still privilege in‑house products over rivals. As these parallel cases around self-preference and Play Store are considered, Google is facing a multi-front regulatory battle that might just reshape its European operations.
To industry professionals, the intersection of technical architecture, content licensing standards, and antitrust enforcement represents a pivotal moment: while AI search evolves toward richer, contextualized answers, its data pipelines-from crawl to model grounding-are increasingly under legal and economic scrutiny. Europe’s push for enforceable licensing, transparent attribution, and fair market access is forcing Google to re‑engineer not just its interfaces, but the very mechanics of how its AI systems ingest, process, and present information.
