A scary question has been hanging over the heads of major online publishers for the past year and a half since ChatGPT was released. What if Google decides to overhaul its core search engine to make generative artificial intelligence more visible, and hurts our business in the process? ?
This question speaks to one of the weakest dependencies in today's online media ecosystem.
Most major publishers, including The New York Times, receive a lot of traffic from people who go to Google, search for something, and click on articles about it. That traffic allows publishers to sell ads and subscriptions, which in turn helps generate the next wave of articles and allows Google to show them to people searching for the next article.
The whole symbiotic cycle has been working more or less successfully for a decade or two. And even when Google announced its first generative AI chatbot, Bard, last year, some online media executives said Google would not put such unstable and unproven technology into its search engine. I consoled myself with the thought that it would never be introduced, or else I wouldn't risk ruining that lucrative search. The advertising business generated $175 billion in revenue last year.
But change is coming.
Google announced at its annual developer conference on Tuesday that it will begin displaying AI-generated answers (called “AI summaries”) to hundreds of millions of users in the United States this week. The company says they will be available to more than 1 billion users by the end of the year.
Powered by Google's Gemini AI technology, the answers appear at the top of search results pages when users search for things like “vegetarian meal prep options” or “Miami day trips.” These provide users with a concise summary of what they're looking for, suggested follow-up questions, and a list of links they can click to learn more. (Users will also see traditional search results, but they will have to scroll further down the page to see them.)
The addition of these answers is the biggest change Google has made to its core search results page in years, and stems from the company's obsession with pushing generative AI into as many products as possible. This may also be a popular feature among users. I've been testing AI Overview for several months through Google's Search Labs program, and I've found it to be generally useful and accurate.
However, it is understandable that publishers are worried. If the AI ​​answer engine works well enough, users won't need to click on links. Whatever you're looking for, it'll show up at the top of search results. And it could undermine the grand “article, we'll give you traffic” deal that underlies Google's relationship with the open web.
Google executives took a positive view of Tuesday's announcement, saying the new AI outline will improve the user experience by “taking the pain out of searching.”
But that effort comes at a cost for a lot of journalism and many other types of online media (fashion blogs, laptop reviews, restaurant listings). Without these media, the Internet would be far less useful. What happens to these websites if Google's AI summaries dry up their traffic? And if large parts of the web disappear completely, what will be left for AI to summarize?
Google clearly anticipated these concerns, and executives prepared a response.
At a press conference this week ahead of Google's developer conference, the company said its tests found that users who were shown an overview of AI tended to conduct more searches and visit a more diverse range of websites. Stated. It also said that links shown in the AI ​​overview received more clicks than links shown on traditional search results pages.
Liz Reid, Google's vice president of search, said in a blog post Tuesday that the company “remains focused on driving valuable traffic to publishers and creators.”
But a careful analysis of these answers reveals that Google is not saying that search traffic across publishers won't decrease. It actually predicts what will happen if Google starts showing AI-generated summaries in his billions of search results per day, and how user behavior will change as a result. Because you can't.
Earlier this year, I wrote about Perplexity. Perplexity is an AI-powered “answer engine” that shows users a concise summary of the topic they're researching, rather than handing them a list of websites to visit. We believed that this experience was clearly better than traditional search engines for some types of searches, and usually could provide more useful information faster.
But I was also nervous, because while testing Perplexity myself, I basically stopped clicking on links at all. It turns out I don't need an AI in a world where it can browse the internet for me and paraphrase what it sees. And I was worried about what would happen if all Perplexity users were like me and had gotten into the habit of relying on AI-generated summaries instead of the original sources.
I have the same concerns about Google's new AI brief, but on a much different scale.
Perplexity is small, with just 10 million monthly users as of February. In contrast, Google has billions of users and accounts for more than 90% of the global search market. If you make a change to a search engine that reduces outgoing traffic by just a few percent, every publisher will feel it.
It's unclear how big an impact Google's AI brief will ultimately have. One analyst firm, Gartner, predicts that traffic to his web from search engines could decline by 25% by 2026. And many publishers are bracing for his double-digit decline in traffic this year.
Perhaps these concerns are exaggerated and publishers have nothing to worry about. But after Tuesday's announcement, Google made it clear that it intends to work it out either way.