Google Unveils AI-Powered Innovations at I/O Conference

Google is set to release a host of new artificial intelligence tools, including an AI assistant designed to help users by performing tasks proactively. The concept of ‘agentic’ AI was a significant focus at Google’s I/O developers conference. One major announcement was the introduction of Gemini Spark, a new AI agent. ‘We are firmly in our agentic Gemini era,’ said Google CEO Sundar Pichai at the event in Mountain View, California. He emphasized that while the potential of such agents is clear, there is still progress to be made in making them user-friendly, secure, and helpful.

Google and its parent company, Alphabet Inc., have invested billions in AI development. Google’s financial plans suggest capital expenditures could reach $190 billion this year. These investments are yielding results; the company’s quarterly earnings indicate robust growth, reflected in an 11% increase in stock value since last month. According to Pichai, the Gemini app boasted 400 million monthly users last year, with current numbers exceeding 900 million, a figure that has more than doubled.

Gemini 3.5: The Latest AI Model

Google is launching a new series of models, starting with Gemini 3.5 Flash, which emphasizes speed and efficiency. The 3.5 Flash is described as the most potent agentic and coding model yet, performing four times faster than some rivals. This model is the default for the Gemini app and Google’s ‘AI mode’. Additionally, Gemini Pro 3.5 is under development internally and expected to launch next month. Each model was trained with improved safety protocols to avoid generating harmful content or erroneously declining safe queries.

Google introduced another model, Gemini Omni, allowing users to create high-quality videos using text, images, videos, or audio as input. These videos are easily edited through interactions with the model. Eventually, Omni will support image and audio creation, though no release date was provided. Google’s Omni model aims to produce more realistic videos, thanks to its understanding of physical forces like gravity and fluid dynamics. Gemini Omni Flash will debut for Google AI Plus, Pro, and Ultra subscribers, with free access on YouTube Shorts and the YouTube Create App.

Videos crafted with Omni include Google’s digital watermark, SynthID. Content credentials will verify if images or videos are AI-created or edited, coming soon to Chrome’s search. AI firms such as OpenAI and ElevenLabs plan to integrate Google’s SynthID technology.

Gemini Spark: A Proactive AI Agent

Gemini Spark, built on Gemini 3.5, will handle boring tasks like organizing meeting notes and emails. It stands out by operating in the cloud, allowing continuous work even when devices are off. Spark seeks user permission for critical actions, such as sending emails or making purchases. It’s currently available to select testers and will enter beta for U.S. subscribers to Google AI Ultra. By summer, it will work directly within Chrome.

Among other announcements, Google updated their smart glasses, which will come in two forms: audio glasses offering spoken guidance and display glasses providing visual data. Audio glasses are expected this fall, with vocal activation and navigation features developed in collaboration with Samsung and eyewear brands. Designs from Gentle Monster and Warby Parker will launch later this year.

Enhanced Search and Shopping with AI

Last year, Google introduced ‘AI mode’ in its search engine, offering conversational responses. This feature has grown in use, surpassing 1 billion monthly users. The new Gemini 3.5 Flash model will now power search, alongside a revamped search box capable of handling extended queries and providing AI-assisted suggestions. Users can search using diverse modalities including text and images. This update starts rolling out in countries where AI mode is available.

A new shopping tool, the Universal Cart, operates across platforms, enabling item addition during browsing. It uses Gemini models to discover discounts, track price changes, and notify users of stock updates. The tool will launch this summer on Google search and the Gemini app, expanding to YouTube and Gmail later.

– Reported by Barbara Ortutay, AP Writer in Oakland, California.