It’s been 17 years since I attended the iconic weekly search quality meeting at Google’s Mountain View campus. During my visit, a diverse group of engineers, product managers, and executives gathered to analyze search query performance and suggest fixes. Back in 2010, Google made 550 changes to its search algorithm in a single year—a figure that felt substantial at the time.
Fast forward to the recent Google I/O developer conference, where the traditional role of search was downplayed by Liz Reid, the head of search. This evolution began two years ago with the introduction of “AI Overview,” which provided summaries at the top of search results while relegating the classic “10 blue links” to secondary status. Reid proclaimed that the most significant transformation in Google’s history now sees users engaging directly with the newest iteration of Google’s AI, Gemini. Queries are now seen as conversation starters, allowing AI to personalize responses using user data, and often generate tailored content on the fly.
This marks a shift from viewing the search box as a portal to the web to an intelligent assistant that produces customized answers. Earlier, Google thrived on interpreting vague search terms. Now, it promotes interaction through conversational prompts with Gemini. The atmosphere at the conference, where Google representatives sported “Ask Me Anything” t-shirts, underscored this change. Instead of engaging with various links, the AI can offer direct answers without redirecting users to other sites.
In a world increasingly dominated by AI, many express discomfort or outright distaste towards its encroachment into our lives—the negative reactions to mentions of AI at public events serve as evidence. Despite this, Google believes that AI search is becoming indispensable, even for skeptics.
I initially resisted the AI Overview feature introduced in 2024, but I’ve since come to recognize its practical advantages for various queries. For instance, when searching for my own past articles, traditional links often proved less helpful than employing straightforward description requests, which yielded immediate results.
According to Google, over a billion people utilize AI Mode monthly, a feature that continues to grow in popularity. Conversations in the search arena vary, with some citing the sentiment that AI-generated responses might signal the demise of traditional web content due to their reliance on existing data.
In conversation with Reid, I asked how they define search in this new context. She emphasized the goal of making information useful and accessible, a mission that traditionally relied on a robust and open web. Reid explained that Google’s current focus is on scraping web pages to gather insights for personalized responses rather than merely linking users to external content.
During the presentation, Google illustrated how a single query could trigger an array of AI agents to craft unique websites dynamically. For example, in response to a question about black holes, AI might create an interactive illustration. However, the source of such information often stems from the hard work of researchers and writers, many of whom remain uncredited in this new framework. Reid contended that while some smaller websites could be negatively impacted, she insists that original content would still retain its audience.
Nevertheless, the potential for AI errors and fabrications remains, an issue Reid acknowledged, though she suggested improvements have been made. Finally, I inquired about the continued existence of search quality meetings like those I attended in the past. Reid revealed that these discussions now occur across multiple sessions, but the focus on human judgement is still paramount. Even if skepticism towards AI persists, its usage continues to rise, with searches reportedly at an all-time high.
