Google's AI training tactics land it in another EU antitrust fight

The European Commission is launching an antitrust probe at Google for allegedly using web and YouTube content to train its AI algorithms while putting competitors at a disadvantage.

According to the Commission, Google may be using content published on the web to fuel its AI Overviews and AI Mode without appropriate compensation to its sources, and without offering publishers an option to refuse inclusion of their content.

As for YouTube content, the Commission said it's similarly worried that Google is using user uploads to train its AI models without appropriate compensation or an option to refuse, noting that content creators "have an obligation to grant Google permission to use their data for... training generative AI models" without any remuneration for their use in that regard.

At the same time, the EC said, Google has banned AI rivals from using YouTube content to train their own models, allegedly giving it a competitive advantage.

"The investigation will notably examine whether Google is distorting competition by imposing unfair terms and conditions on publishers and content creators, or by granting itself privileged access to such content," the Commission said in a statement. "If proven, the practices under investigation may breach EU competition rules that prohibit the abuse of a dominant position."

Google disagrees, arguing in the same manner as other US tech giants targeted by Europe's strict competition rules that the Commission's move will put the brakes on innovation.

"This complaint risks stifling innovation in a market that is more competitive than ever," a Google spokesperson told The Register. "Europeans deserve to benefit from the latest technologies and we will continue to work closely with the news and creative industries as they transition to the AI era."

Google further explained how it believes it's acting correctly under EU law, telling us that its Google-Extended crawler token gives publishers control over what its bots can crawl, and reiterating that it respects robots.txt directives.

That said, crawlers have been known to behave badly when facing directives from websites, and restricting Google bots from indexing websites can have a direct effect on search rankings - one of the very things the Commission expressed concerns about in today's antitrust investigation announcement.

As for YouTube content, Google told The Register that it's been working with creators to protect their businesses through its likeness detection tool that flags creators when their image is being used in AI-generated content. We note that doesn't address any of the EU's concerns that content creators aren't being given the option to avoid having their videos ingested to train Google AI.

This antitrust investigation comes less than a week after the EC opened another into Meta related to AI restrictions. In that case, EU regulators are investigating changes to the company's WhatsApp messaging platform that banned rival AI chatbots. The move has led to both ChatGPT and Microsoft Copilot from exiting WhatsApp, while Meta's own AI remains available.

Meta defended itself by calling the EC's claims "baseless," saying that the change was made because WhatsApp wasn't built to handle the load placed on it by the presence of multiple AI chatbots accessing the platform via a single API. ®