If you haven’t heard yet, the latest round of Google I/O announcements involve a mixed bag of updates but the primary focus has remained steadily on AI. The first day of the Google I/O event held on May 20-21, 2025, at the Shoreline Amphitheatre in Mountain View, California, included a showcase of the company’s primary developments, including the Gemini 2.5 AI model that moved out of preview mode.
The Google AI upgrades extend beyond just Gemini to the Search function, Gmail, and other services, even opening the doors wider for third-party apps to integrate its tools. While we expected Google I/O to feature more Android updates, we did receive a healthy dose of them earlier this month, covering the Material 3 expressive design language coming to Android 16 later this year.
For now, let’s break down everything we need to know about the Google AI upgrades.
The Google I/O 2025 event has opened the floodgates on AI.
Google I/O 2025 Announcements—What’s Cooking In the Google Labs?
The Google I/O highlights make it obvious that the company is doubling down on its investment into artificial intelligence. The AI obsession across the tech industry is as overwhelming as it is unpleasant, but we have no choice but to shrug it off with a “it is what it is,” and move on to discuss the advancements as they have been laid out. The Google Assistant that we’re all familiar with but rarely use is being replaced by Gemini, and she isn’t arriving without her bag of updates.
Google Gemini Updates
Gemini Live is the latest advancement to the AI’s capabilities, combining inputs from the camera and voice commands to perform different actions like making calls and searching the web. We saw Project Astra showcased last year, and its capabilities are making its way to the tool across iOS and Android devices. The multimodal search capabilities are the service’s most significant feature, but Gemini 2.5 AI model is capable of quite a bit more.
Google’s universal AI assistant was updated with smarter, more proactive features, including access to personal information and the ability to speak up proactively to communicate updates. The company’s LearnLM tool has been infused directly into Gemini 2.5, making it a top choice for learners according to Google.
Gemini 2.5 Pro was also updated with enhanced coding capabilities, and was released earlier on May 6, 2025. Sundar Pichai’s Google updates made a mention of the fact that the Gemini app has more than 400 million active monthly users, which is an impressive number no matter how you look at it.
Gemini Nano for Third-Party Apps
Google opened APIs for Gemini Nano, its on-device large language model, allowing developers to integrate it for tasks like transcription, summarization, and writing assistance. As developers familiarize themselves with the APIs, we could see greater AI integration across all the apps we love and use on a daily basis.
New Generative AI Tools Mentioned in the Google I/O 2025 Announcements
Google has an entire arsenal of AI tools to offer beyond relying exclusively on Gemini. The company announced Veo 3, its latest video generation model where users can now seek out clips with audio as well. The older Veo 2 model was updated with better camera controls, outpainting, and object eraser for videos generated using the tool. Imagen 4 was added to the Gemini app, Whisk, Vertex AI and across other apps, for enhanced image generation.
Also relevant to the Google I/O 2025 recap, the company revealed the Flow AI tool exclusive to Google AI Pro and Ultra plan subscribers. The tools combine the capabilities of Veo, Imagen, and Gemini, allowing for the creation of beautiful scenes with a simple prompt.
For music lovers and creators, Google explained its Music AI Sandbox tool powered by Lyria 2, for exploring the use of AI in music generation. The tool is available for creators through YouTube Shorts—a feature Google constantly tries to push despite its unappealing presence—and enterprises in Vertex AI.
Google AI Mode Search
Next, we have to talk about the tech giants’ next attempt at making AI searches more relevant. The AI mode, a new tab in Google Search, was rolled out to all US users, moving it out from the preview phase in Google Labs to broader availability.
This brings a multitude of new features such as a shopping tool with a virtual try-on function using a photo, agentic checkout, price watcher, and the ability to compare ticket prices and create custom charts or graphs for topics like finances or sports. Despite the chaos of Google’s AI Summary tool constantly churning out incorrect responses, we have to admit the shopping function is particularly interesting this time around.
This Google AI mode search is exceedingly more advanced than the AI search summary we’ve been submitted to for the last year, and could be a turning point in bringing more users on to the side of AI.
Google AI Ultra Plan—Setting the Price of Premium AI
A large number of these AI tools are free to use but for the more exclusive experiences, Google has an AI Ultra Plan for $249.99 per month, which will increase the limits on AI use and provide access to experimental tools like Project Mariner. This refers to a new browser extension to automate keystrokes and mouse clicks, which is a suspicious tool but a useful one all the same.
With the Google AI Ultra Plan, you also get access to Gemini 2.5 Pro, Flow, higher limits on Whisk, NotebookLM, an individual YouTube premium plan, and 30TB of storage. Gemini is also better integrated into Gmail, Docs, Vids, and other Google apps.
First-time buyers can get 50% off for their first three months so if you’re committed to mastering Google’s AI offerings, now is the time to buy in. The existing Google AI Pro plan is also available to consider if you want a more affordable subscription, but it does come without a few of the features.
Say Hi to Jules, Google’s “Asynchronous, Agentic Coding Assistant”
Google introduced Jules to Google Labs last year, but the coding agent has since evolved to become a more efficient coding assistant than before. “It clones your codebase into a secure Google Cloud virtual machine (VM), understands the full context of your project, and performs tasks,” according to the company, and it aids with fixing bugs, writing tests, and generally perfecting your code.
Google I/O Android XR Updates
Setting aside the artificial intelligence for a moment, the Google I/O 2025 announcements also extended to the company’s showcase of its Android XR platform for headsets and smart glasses, featuring a prototype developed with Samsung (Project Moohan), set to launch later in 2025.
We got a glimpse of the glasses last year, but Google brought back attention to the prototype glasses, which feature a discrete camera and display as well as a microphone and speakers. Through the Gemini integration on the device, you get real-time updates, including displaying directions or answers to queries about your surroundings.
Google is partnering with Gentle Monster and Warby Parker to bring these glasses to life with its Android XR platform. If you thought glasses with cameras and hidden assistants were exclusively reserved for secret agents like in the movies, Google is planning to change that really soon, in ways that are more advanced than the Meta Ray-Bans.
Final Mentions in our Google I/O 2025 Overview
The Google I/O 2025 highlights we just mentioned merely scratch the surface of everything the business is up to right now. There’s also Google Beam, which is the company’s first 3D video communication platform, and the upgrades to the SynthID detector, which is a portal for identifying AI generated content. The Gemini Code Assist tool is advancing by leaps and bounds and we haven’t even started on all the experimental work the business has involved itself in, research technology that far exceeds what we have today.
Like it or not, the scale of work being done at Google is more expansive than the capabilities of any other organization, and it’s apparent when you consider just how many business avenues it is involved in. Google is paving the way for AI to become the most integral part of our daily lives, and the Google I/O 2025 announcements are just a hint of what that means for the future.
Have any insights to share with us on the Google I/O event? Let us know what you think. Subscribe to Technowize for more insights into the latest on-goings in the world of technology.