Google Revolutionizes Search with AI Overviews

Just over a year ago, Google announced its vision for the future of search powered by artificial intelligence (AI). Now, that future has arrived. Google is unveiling its “AI Overviews” feature, previously referred to as the Search Generative Experience (SGE), to users in the US, with a global rollout planned soon. This means that billions of Google users will soon encounter AI-generated summaries at the top of many of their search results. And this is just the start of AI transforming search.

Liz Reid, Google’s new head of Search, who has been immersed in AI search development, explains, “Generative AI allows Google to handle much of the searching for you, streamlining the process so you can focus on achieving your objectives or exploring the topics that excite you.”

Reid highlights numerous features that make this possible, all revealed at Google’s recent I/O developer conference. Among them are the AI Overviews that provide a concise answer to your query along with links for further reading. Additionally, a new Lens feature lets users search via video capture. There’s also an innovative planning tool that can automatically generate trip itineraries or meal plans based on a single query. Furthermore, an AI-driven method organizes the results page, offering suggestions tailored to your needs, such as restaurants for different occasions without explicitly being asked.

This transition represents a comprehensive AI enhancement of search functionality. Google’s Gemini AI interprets user queries regardless of whether they are typed, spoken, or visual. A specialized Gemini model then summarizes the web to produce an answer and even designs the results page.

Google leverages AI to both populate and structure your search results page. Image: Google

However, not all searches necessitate extensive AI involvement, as Reid points out. Simple queries, like searching for “Walmart” to navigate to walmart.com, don’t benefit much from AI. Where Gemini proves most useful is with more complex queries that would typically require multiple searches or might not be addressed by conventional search methods.

Reid uses local search as an example—often a complex task due to numerous similar listings and reviews. With Gemini, users can find specific results, such as the best yoga or pilates studios in Boston rated over four stars within a half-hour walk of Beacon Hill. The AI can also integrate information about the best offers for first-timers by pulling data from Google’s Knowledge Graph and the web.

The integration of the Knowledge Graph with AI is crucial for Reid and her team. While some search functions, like checking sports scores, are already optimized, Gemini ensures accurate results regardless of query phrasing. “We’re focusing on expanding the range of queries that can successfully return sports data,” she says, emphasizing the importance of reliable data.

While not every search will feature an AI overview, many will.

Accurate data is essential for any search engine. Reid explains that one motivation for the new Gemini model was to prioritize accuracy over creativity. “We aim to tilt towards factuality,” she remarks. AI Overviews may sacrifice some charm for accuracy, though no model is flawless, and Google will undoubtedly encounter issues with inaccurate summaries.

The arrival of AI in search has raised concerns, particularly regarding services like Perplexity and Arc, which crawl and summarize the web without proper attribution. Reid acknowledges the challenge but believes the new search method will increase clicks to valuable sources. She argues that while low-quality content might suffer, high-quality websites offering unique insights and expertise will benefit as users look for deeper understanding.

Over the past decade, Google has been evolving from a simple keyword-search box into an intelligent assistant capable of understanding and responding to queries in various forms. “By enriching the search experience and allowing natural queries, we attract more questions,” Reid concludes. This transformation aims to enhance user interaction with the internet, promoting less typing, fewer open tabs, and more conversational engagement with the search engine.