Google Breaks Through with a Series of New AI Features

On May 20, the American technology corporation Google officially launched a series of breakthrough improvements revolving around artificial intelligence (AI), focusing on "reshaping" the Google Search search engine with a completely new AI feature - AI Mode.

During the opening keynote at the Google I/O 2025 annual conference held in Silicon Valley, CEO Sundar Pichai called AI Mode “the ultimate search engine.”

He noted that this AI feature is the result of “decades of research” and is a huge step forward from the previously launched AI Overviews, which now has more than 1.5 billion users worldwide.

With AI Mode, users can ask long, complex questions, follow-up queries, and get answers with advanced reasoning, while AI Overviews simply displays AI-generated answers above traditional links.

According to Ms. Liz Reid - Director of Google Search, AI Mode has now been deployed in the US and is a powerful tool with complex, multi-modal reasoning capabilities and allows users to dig deeper into queries.

Also at this event, Google announced a series of other upgrades in its AI ecosystem, with the highlight being the Gemini 2.5 Pro model integrating a new inference feature called “Deep Think.”

Deep Think allows the AI ​​model to evaluate multiple possible answers before responding, significantly improving performance on selected benchmarks. While the feature is currently in the testing phase for accuracy and safety, the potential of Deep Think in the future is already evident.

Additionally, Google launched a premium service package Gemini Ultra for $249.99/month.

This service package unlocks Google's most advanced AI tools, including: the next-generation Veo 3 video editing app that allows you to create videos from text or images, with audio, background music and even dialogue; the Imagen 4 app - a super-detailed AI image creation tool, 10 times faster than the previous generation, supporting both realistic and abstract styles; the Gemini Live app - a real-time conversation interface that can handle input from the camera or phone screen, interacting directly with Google Maps, Calendar or Tasks.

Google also introduced AI Agents – virtual assistants that can perform online tasks independently on Chrome and the Gemini app. These assistants are designed to automatically search for information, compare prices and shop when the right deals come up.

Google Workspace users will also be supported by a series of new AI features such as: smart Gmail inbox cleaning, personalized replies, automatic document writing in Google Docs or creating summary videos via Google Vids.

According to Mr. Pichai, the Gemini application now has more than 400 million monthly active users, a sharp increase from the 350 million figure stated in an internal report last March./.

(Vietnam News Agency/Vietnam+)

Source: https://www.vietnamplus.vn/google-cai-tien-dot-pha-voi-hang-loat-tinh-nang-ai-moi-post1039782.vnp