Google’s New AI Developments: A Glimpse into Gemini

Google is set to introduce a range of new AI-driven tools and systems, including an assistant that anticipates user needs by performing tasks autonomously. This concept of ‘agentic’ AI was prominently featured at Google’s annual developers conference, Google I/O. Among the notable announcements was Gemini Spark, part of the company’s latest AI innovations.

Agentic Era at Google

CEO Sundar Pichai emphasized the significance of this era during the conference held at Mountain View, California, mentioning the potential and early stages of making AI agents highly efficient. Google and Alphabet Inc. have heavily invested in AI, with their capital expenditures expected to hit $190 billion this year. The financial outcomes are promising, with recent quarterly earnings reflecting robust growth. The stock has seen an 11% rise following the earnings report.

Gemini App User Growth

According to Pichai, the Gemini app boasted 400 million monthly active users last year, a figure that has now more than doubled, reaching over 900 million. The app’s evolution continues with the rollout of Gemini 3.5, starting with Gemini 3.5 Flash, which prioritizes speed and serves as a default for the Gemini app and Google search’s ‘AI mode’. The upcoming Gemini Pro 3.5, currently in internal use, is slated for release next month. Safety training in Gemini 3.5 has been enhanced to reduce harmful outputs and address queries more effectively.

Gemini Omni and Content Creation

Gemini Omni, a new model, allows users to create sophisticated video content with various inputs such as text, images, and audio. These videos, made through conversational interaction, are claimed to be more realistic due to Omni’s advanced understanding of physics. The initial version of Omni, called Gemini Omni Flash, debuts for Google AI Plus, Pro, and Ultra subscribers. YouTube Shorts and the YouTube Create App will also offer this feature without cost, and videos will carry Google’s SynthID watermark for authenticity.

Advanced AI Tools and Inclusivity

Google is also collaborating with AI companies such as Open AI and Eleven Labs to expand the use of SynthID technology. The Gemini Spark platform, powered by Gemini 3.5, is designed to efficiently manage tasks like sorting emails and creating summaries from notes. Unlike standard chatbots, it operates in the cloud, continuously working even when devices are inactive. Spark requires permission for tasks like emailing or purchasing, ensuring user control.

Innovative AI Wearables

Adding to Google’s diverse AI announcements are two smart glasses types: audio glasses for auditory assistance, and display glasses for visual information. Google, in partnership with Gentle Monster and Warby Parker, aims for a release as part of their collections. The audio glasses, arriving this fall, feature hands-free interaction via voice or touch commands.

Enhancing Search and Shopping with AI

The conference highlighted last year’s introduction of ‘AI mode’ in Google search, enhancing user interaction with conversational responses. Usage has surged, surpassing 1 billion monthly active users, according to Liz Reid, head of search. The Gemini 3.5 Flash will replace the default search model, featuring an adaptive search box capable of handling extensive queries and suggesting AI-powered content. This rollout begins globally where AI mode exists.

Moreover, Google introduced the Universal Cart tool, promising a smart shopping experience. It integrates across services, searching for products and alerts on deals or restocks. Users will access this tool through search and the Gemini app, followed by expansions to YouTube and Gmail.

Note: Barbara Ortutay of the Associated Press contributed to this reporting.