Then again, to describe those billions as "users", as parent company Alphabet did when announcing its quarterly earnings last week, is perhaps disingenuous. No one consciously uses AI Overviews - it's just there when users perform a regular search on Google, something billions of them have done several times a day for two decades.
That's one key advantage Google has over its competitors: people already associate the service with finding things out. The company has every right to capitalise on that reputation, one it built off the back of genuine innovation and quality though, admittedly, it was later solidified with illegal multibillion-dollar deals to prevent competition.
Google's second advantage with AI Overviews, however, warrants further scrutiny. Like other generative AI tools, the feature draws heavily from content that Google does not own but is available on the open web. Summarised answers are synthesised from one or more sources into a rewritten piece of information.
This is about taking away Google's cheat code. "Google still thinks they're special and that they don't have to play by the same rules that the rest of the industry does," Matthew Prince, CEO of Cloudflare, said in an interview last week. His company recently launched a tool that would allow publishers to set up a "pay-per-crawl" model for AI use. It works on crawlers from OpenAI, Anthropic, Perplexity and most others - but blocking Google AI would, again, mean blocking a site from Google's search engine.
In Google's defence, the launch of AI Overviews was a move spurred not by a desire to crush the economics of web - which has driven its entire business - but to stop its users from deserting the company in favour of AI chatbots. "The consumer is forcing them," Wells Fargo analyst Ken Gawrelski said. Google was more than satisfied with the status quo, Gawrelski told me, which is partly why the company was beaten to market by smaller AI firms that didn't need to worry about protecting an existing revenue stream.