According to Fortune, European Union antitrust regulators opened a formal investigation into Google on Tuesday, probing whether the tech giant illegally used online content to train its artificial intelligence models. The European Commission is specifically examining Google’s use of content from web publishers and material uploaded to YouTube for AI services like AI Overviews and AI Mode. Officials are concerned Google gained an unfair edge over rivals by using this content without paying creators or allowing opt-outs, potentially breaching EU competition rules. The investigation, which has no set deadline, could result in sanctions including a fine worth up to 10% of Google’s annual global revenue. EU spokeswoman Arianna Podesta stated the Commission is “agnostic” about company nationality and has informed U.S. authorities about the probe.
The core issue is consent and control
Here’s the thing: this isn’t just about scraping data. Practically every AI company does that. The EU’s argument seems to hinge on market power. They’re asking if Google, because it owns the dominant search engine and YouTube, can impose unfair terms or grab privileged access to content in a way that a smaller startup simply couldn’t. It’s the “self-preferencing” argument we’ve seen before, but applied to the AI training data pipeline. The worry is that Google can use YouTube’s terms of service—which users agree to when they upload—as a blanket license to train its models, while simultaneously locking rival AI developers out of that same data trove. That’s a powerful dual advantage if true.
Google’s defense is predictable and wobbly
Google’s statement, that this “risks stifling innovation in a market that is more competitive than ever,” is the standard Big Tech playbook. But it feels especially hollow right now. More competitive? In search, maybe not so much. In AI, they’re famously playing catch-up to OpenAI. This investigation suggests regulators think the way Google is trying to catch up might be the problem. They’re not attacking the innovation itself, but the methods used to fuel it. And let’s be real, saying you’ll “work closely with the news and creative industries” rings pretty flat when you’re being investigated for potentially not doing that in the first place.
This is part of a much bigger EU crackdown
Look, this isn’t an isolated case. The Commission just opened a probe into WhatsApp’s AI data use and recently fined Elon Musk’s X. They’re using their classic competition rulebook, not just the newer Digital Markets Act (DMA), which is interesting. It shows they’re reaching for every tool they have. The political tension is obvious too, with the mention of Trump administration complaints. The EU is basically saying, “We don’t care where you’re from, if you do business here, these are the rules.” But you can bet this will be framed as another example of Europe targeting American tech success.
The stakes are insanely high
A fine of up to 10% of global revenue is no joke. For Google, that’s tens of billions. But the real cost is precedent. This investigation is digging into the foundational practices of the AI era: how models are trained. If the EU forces a new model of consent and payment for training data, it doesn’t just change Google’s game. It changes the game for everyone. Will AI companies need to license every snippet of text, every video frame? That could slow progress to a crawl or entrench giants who can afford the licensing fees. It’s a messy, ugly fight with no easy answers. And honestly, it’s one we probably should have had before the AI race went into overdrive.
