The New York Times argued yesterday that Google’s AI search features are starving the open internet, which depends in part on user traffic to sustain itself. I worry a lot about the directions AI could go in the future, but I find myself deeply skeptical of this particular fear.
The claim, as I understand it, is that users ask a question or search for a topic on Google, an AI answers their question or summarizes the topic, and the user goes on their way, not having clicked on any websites.
This probably seems tragic from the perspective of websites optimizing for clicks, but it sure seems like an improvement to me. When I Google “how to replace my car’s air filter”, I don’t want to hack my way through a jungle of AutoZone ads and reddit threads, I want to know how to replace my car’s air filter. If Google’s AI is helpful enough to tell me the answer so I can get on with my life, good.
The parts of the internet that rely most heavily on clicks from Google are also the parts that depend largely on ad revenue. And those parts kind of suck. If AI Mode forces sites to find a business model that doesn’t rely on a desperate bid for human eyeballs, I won’t mourn the passing of that era.
I am more sympathetic to the idea that sites like Wikipedia are seeing less traffic as a result of AI summaries. But Wikipedia relies on donations, not ads. If I want Wikipedia to stay around (and I do! It’s a great website!), I can donate to Wikipedia.
I can’t speak for everyone, but “Ask a plain English question, get a plain English answer with citations” is much closer to the way I intuitively want Google to work than “Ask a plain English question, get a bunch of links that are sort of related to it, click them and hope.” In that sense, AI Mode feels like a natural evolution of Google into the thing we always really wanted.
Of course, AI Mode isn’t always accurate. But neither are the websites that it pulls from. In my experience, it’s usually good enough for a first-pass, and I can always dig deeper when I need to.