Google’s New AI Innovations: Gemini Spark and More

Google is set to introduce a host of new artificial intelligence tools, highlighted by an AI assistant that performs tasks for users. The focal point of this year’s Google I/O conference, ‘agentic’ AI, was emphasized through the unveiling of Gemini Spark, the latest AI agent. ‘We are firmly in our agentic Gemini era,’ stated Google CEO Sundar Pichai at the event in Mountain View, California. While early in development, the potential of these agents is notable, Pichai remarked.

Google and parent company Alphabet Inc. have invested significantly in AI. Their chief financial officer mentioned potential capital expenditures reaching up to $190 billion this year. This investment appears fruitful, reflected in Google’s rising stock, which increased by 11% since the last earnings report. The Gemini app, a key player in Google’s AI lineup, saw its monthly active users soar from 400 million to over 900 million in just a year.

Gemini 3.5 Launch

Google is rolling out its latest models, Gemini 3.5, starting with Gemini 3.5 Flash. This model prioritizes speed and is claimed to be four times faster than some competitors. Gemini 3.5 is now the default for Google search’s AI mode. Internally, the company is preparing to launch Gemini Pro 3.5 next month.

Gemini 3.5 was developed with enhanced safety training, reducing harmful content generation. Google also unveiled Gemini Omni, which enables video creation using varied inputs like text, images, and audio. Omni also supports editing through conversation. Videos created with Omni are designed to be realistic by understanding physical forces. Gemini Omni Flash is available for Google AI Plus, Pro, and Ultra subscribers, and all Omni videos will feature Google’s SynthID digital watermark.

Google announced SynthID’s adoption by AI companies like Open AI and Kakao, extending its technology to more AI-generated content. SynthID, with content credentials verification, helps determine AI-created images and videos.

Gemini Spark and Smart Glasses

Gemini Spark, powered by Gemini 3.5, streamlines tasks by sorting through meeting notes, emails, and creating documents from summaries. As a cloud-based agent, Gemini Spark continues operations even when devices are inactive. The system seeks user permission before performing significant tasks, like sending emails. Select testers can access the agent, and a U.S. beta rollout is planned. By summer, Gemini Spark will integrate with Chrome.

Google also updated its smart glasses initiative. Two variations are in development: audio glasses for spoken guidance and display glasses for visual information. The audio glasses, debuting this fall, interact with Gemini for navigation, communication, translations, and more. Google partnered with Samsung and brands Gentle Monster and Warby Parker to produce these glasses, set for release in upcoming collections.

Enhancements in Search and Shopping

Google’s ‘AI mode’ in search, introduced last year, offers conversational responses to queries. This feature now surpasses 1 billion monthly users with query numbers doubling quarterly. The enhanced search model Gemini 3.5 Flash will become the search default. A significant update to the search box offers adaptation to longer queries with AI suggestions, facilitating multi-modal search using various media.

Google’s new Universal Cart aims to redefine online shopping. This tool consolidates shopping across Google services, seeking deals, tracking prices, and notifying availability. Available initially in search and the Gemini app, Universal Cart expands to YouTube and Gmail later.