This illustrated photo shows the Google logo displayed on a laptop screen. (Photo illustration by Rafael Henrique/SOPA Images/LightRocket, Getty Images)
Google announced Tuesday that it will introduce AI-generated answers to online queries by users in the United States, in one of the biggest updates to its search engine in 25 years.
“I'm excited to announce that this week we're launching a completely new experience, AI Overviews, to people in the U.S.,” Google CEO Sundar Pichai said at an event in California. Stated.
This feature will soon be available in other countries, he added.
With this change, many of Google's search results will now display an AI “overview” at the top of the page, before rolling out general links and features.
The search engine's AI answers, generated by Google's Gemini AI technology, provide a one- to two-paragraph explanation with links to the online sources that provided the information.
“From research to planning to brainstorming, you can ask us anything you care about or need to accomplish, and we'll take care of it for you,” said Liz Reid, head of Google's search team. said.
The change appears to be a response to growing pressure from AI-powered search engines like Perplexity and repeated rumors that ChatGPT creator OpenAI is building its own AI search tools.
Creators and small publishers are nervous about this change, worried that users will stop clicking on websites to find information.
Research firm Gartner predicts that search engine traffic to the web will decline by 25% by 2026 due to the introduction of features such as AI bots.
Google has pushed back against suggestions that ChatGPT-style chatbots could impact its business.
“We found that with AI Overview, people used search more and were more satisfied with search results,” said Reid.
“Rather than splitting your questions into multiple searches, you can ask your most complex questions all at once, keeping in mind all the nuances and caveats.”
“Intuitive and useful”
Google's announcement was made as part of an AI-focused keynote opening Google's annual I/O developer conference in Mountain View, California.
Reid said the company will also soon begin testing the application of AI to searches based on video content as a query source.
Multimodal queries like this are one of the highlights of GPT-4o, OpenAI's flagship model that can generate content and understand voice, text, and image commands, which OpenAI released on Monday. This is an update.
OpenAI's technology updates have proven to be highly conversational, helping people tell jokes, write songs, and teach algebra to students.
Google, like OpenAI, showed off a demo in which staff asked its AI to recognize its surroundings through smartphone video cameras and other assistant-like skills.
According to Sissie Hsiao, general manager of Gemini Experiences and Google Assistant, the feature is designed to be “conversational, intuitive, and helpful.”
“Collaborate with the most intelligent and personalized Gemini ever.”