Google Unveils Major AI Initiative at I/O 2026

May 20, 2026, 9:30 am | Read time: 4 minutes

At the Google I/O 2026 developer conference, the company unveiled numerous new AI features. At the center is Gemini 3.5. The new model is set to power many Google services in the future and form the basis for new AI functions. This involves not only web search but also apps, shopping, and personal assistants.

Google describes Gemini 3.5 as the company’s fastest and most reliable AI model to date. It starts with the “Gemini 3.5 Flash” variant. A pro version with enhanced understanding and higher accuracy is to follow later. The model is used, among other things, for the AI mode of Google Search and for developer services like Google AI Studio and Android Studio.

New Features with Gemini Spark

With “Gemini Spark,” Google is also announcing a personal AI assistant. The assistant is intended to connect various Google services and help users better organize their digital daily lives.

Spark accesses content from different applications, including mail, calendar, or notes. The assistant is supposed to summarize tasks, create reminders, or monitor important deadlines. According to Google, Spark can also monitor emails from certain senders and automatically create summaries.

Also interesting: Users will soon be able to control Gemini’s cognitive load themselves

Additionally, the assistant can be controlled by voice. Google also plans for users to teach the system additional skills. Spark is currently still in the testing phase and will initially launch in the U.S. for subscribers of the Google AI Ultra program. When the service will come to Germany is still open. Google promises that users can determine which apps the assistant can access. Actions will only be executed with explicit consent.

Gemini Omni to Improve AI Videos

Google is also planning major innovations in AI-generated videos. For this, the company is introducing “Gemini Omni.” The system combines video editing, generation, and contextual understanding.

Google showcased examples where existing video footage was significantly altered. For instance, the AI transformed a stone environment into a scenario of soap bubbles. According to Google, Omni uses extensive datasets on materials, physical properties, references, and various styles for this.

The company also sees applications in the education sector. There, Omni is to explain complex content using interactive and animated videos. The feature will initially launch as a “Flash” version for users of Google AI Plus, Google AI Pro, and Google AI Ultra. The technology will also be available for free in YouTube Shorts and YouTube Create.

AI Content to Become More Recognizable

With the increasing number of AI-generated videos, there is also a growing demand for labeling such content. Google announced that content from Gemini Omni will be marked with the “SynthID” watermark.

Additionally, according to Google, other companies like Nvidia, OpenAI, and Kakao also want to use this labeling.

With Gemini Intelligence Google Announces Largest Android Update in Years

Privacy Alert Google AI Apparently Reads WhatsApp Messages

Google Search to Focus More on AI

Google Search is also receiving new AI features based on Gemini 3.5 Flash. The so-called AI mode is expected to deliver faster and more precise answers. The company also plans changes to the interface. The search bar will become larger and better handle natural language. Additionally, new search agents will provide personalized content and individual feeds.

With “Personal Intelligence,” Google also connects content from Gmail, Google Photos, and Google Calendar with search queries. The prerequisite is that users are logged in.

Also interesting: Revealed! AI for businesses more expensive than human staff

Shopping Offensive with Universal Cart

Google is also increasingly relying on AI for online shopping. The company announced several new systems for this. The “Universal Commerce Protocol” is intended to better connect retailers, suppliers, and service providers. At the same time, the system is supposed to simplify payment processes and help users find suitable offers more quickly.

Additionally, the “Agent Payment Protocol” is coming. According to Google, it is intended to securely connect customers, retailers, payment services, and AI systems. With “Universal Cart,” Google is also planning an automatic shopping function. Users will be able to add products to the cart and set desired prices. If the price falls below the specified value, the AI agent can automatically complete the purchase.

The launch of Universal Cart is planned for the summer in the U.S. It is still unclear if and when the features will also appear in Europe.

This article is a machine translation of the original German version of TECHBOOK and has been reviewed for accuracy and quality by a native speaker. For feedback, please contact us at info@techbook.de.