Image credit Google
Google I/O 2025
“Beyond information, to intelligence.”
Google held their annual Google I/O event starting on May 20th, where they unveiled their newest platform announcements for the year. Artificial Intelligence advancements were front and center throughout the whole event, touching every area of the Google development ecosystem, from user-facing implementations like the Gemini assistant, developer-facing implementations including updates to developer tools and new APIs for integrating with AI services, and behind-the-scenes implementations powering conversational and mixed-reality experiences.
Here are five key focus areas that technologists and digital strategists should be planning for in their development roadmaps over the coming year.
The latest version of Android is right around the corner, tentatively scheduled to release in the second half of 2025. Android developers should update their apps to target Android 16 and thoroughly test them to ensure they’re compliant with the new SDKs. Here’s a breakdown of some of the most significant features developers should be aware of:
Live Updates
Similar to iOS’s Live Activities, Android 16’s Live Updates are a new class of notifications for tracking the progress of ongoing events. These notifications allow for frequently-updated information to be displayed on the Lock Screen and status bar, and are ideal for use cases like rideshare pickups, food delivery alerts, and navigation.
More info
Material 3 Expressive
Google has updated its Material design system to Material 3 Expressive, designed to help apps feel more natural and engaging. It introduces new emotional design patterns aimed at boosting user engagement and usability, and a combination of interactive physics-based animations and haptics make the Android experience feel more responsive and lively. Material 3 Expressive also gives users the ability to personalize their experience with their own color settings and other preferences.
More info
Edge-to-edge Layouts Now Required
Prior to Android 16, developers had the option of opting out of implementing edge-to-edge layouts. But starting with Android 16, support for edge-to-edge layouts is now required. Developers should ensure their apps correctly implement edge-to-edge layouts in their designs and implementations going forward.
More info
Changes to Back Navigation
Android’s back-navigation APIs have gradually changed over the past few versions, including the addition of predictive back gestures that give users glimpses into where they’re navigating when interactively swiping backwards in the navigation stack. Starting with Android 16, predictive back gesture animations will now be enabled by default, making this an important time to review your implementations of back-navigation APIs and update or opt out as appropriate.
More info
Other Noteworthy Changes
Other Android 16 behavioral changes include updates to the elegant font APIs, health and fitness app permissions, safer intents initiative, and a variety of others. For a more comprehensive roundup of the complete set, see https://developer.android.com/..., and be sure to update your apps accordingly for your own relevant features.
Adaptive apps are a design paradigm shift in the Android ecosystem. They emphasize allowing users to smoothly run apps across a variety of devices in any orientation, window size, and aspect ratio. Users on larger-screen devices like tablets will expect to be able to take complete advantage of their screen real estate, and adaptive apps enable this. And adaptive apps also run on devices like foldables, desktops, cars, TVs, and mixed reality wearables.
If you're already targeting larger-screen devices and haven't implemented the adaptive app model, schedule this into your development roadmap soon. Apps will need to start adhering to the new paradigm once they target Android 16, with the option to opt out if needed available in 2025. Starting in 2026, no opt-outs will be allowed.
And if you're not currently targeting larger-screen devices, now is a great time to start! It's easier than ever to build interfaces that smoothly scale to fit a wide variety of devices and window sizes. This will also put you into a great position to easily expand onto other devices including foldables, cars, TVs, and mixed reality platforms.
Smaller-screen devices and games are currently exempt from these requirements. However, it's not a stretch to expect they may also need to comply at a future date. And by correctly implementing adaptive app interfaces for smaller-screen devices, expanding into larger-screen devices will require minimal additional effort without needing to support multiple completely different layout paradigms per device class.
Google is also working to make it as easy possible to implement adaptive interfaces with SDK updates to the Compose adaptive layouts library, window size classes, Jetpack Navigation 3 (currently in alpha), and enhancements to the development tools for previewing and testing adaptive layouts (such as Journeys with Gemini in Android Studio that allow you to define tests using natural language).
Android is also expanding into the third dimension with Android XR, Google’s platform for building immersive experiences for glasses and headsets. The platform beautifully leverages Google’s AI-powered Gemini assistant to deliver an intuitive, hands-free experience on Android XR wearable devices. Android XR promises to deliver powerful and natural experiences like navigation, live language translation, photos, and a multitude of other app experiences perfectly suited towards the wearable device landscape.
Google revealed a brief glimpse into the first Android XR devices slated to hit the market this year, including Samsung’s Project Moohan, in a headset form factor. And even more interestingly, they also demonstrated a live prototype of their new glasses-based device on stage. With a clear lineage stemming from the original Google Glass hardware from around 10 years ago, it’s exciting to see the technology now reaching its full potential. The glasses form factor also clearly sets Google’s device apart from Apple’s Vision Pro headset form factor, and it’s easy to imagine this device smoothly integrating into wearers’ daily lives.
Android XR devices will be available later this year, but developers can start preparing apps for them right now. The platform supports native Android-based development with the Android Jetpack XR SDK, allowing developers to build apps with the same development languages, toolkits such as Jetpack Compose, and Material Design paradigm that Android developers already know and love. Existing Android apps can even run directly on Android XR with minimal development effort. And other tools such as Unity, OpenXR, and WebXR are also supported, giving developers a wide array of options for bringing their XR experiences to life.
Model Context Protocol (MCP) is an open protocol introduced in November 2024. MCP is designed to standardize how AI models (like Gemini) integrate and interact with external tools, data sources, and services. AI models, especially agentic ones that can perform tasks, often need to access external information or trigger actions in other systems (e.g., search the web, manipulate a database, interact with a calendar, access a GitHub code repository, etc.).
In short, MCP implementations are like apps for agents. Brands should consider a strategy for how they might want to expose their content and functionality so that they can be front-and-center and relevant as the AI ecosystem expands.
Google announced that their Gemini SDKs now support MCP. Additionally, developers can now build AI agents leveraging Gemini 2.5's advanced reasoning capabilities with new tools like URL Context. This feature enables the model to extract context from web pages using just a link.
Using the Gemini Live API, developers can now build agentic applications that understand and speak 24 different languages. The Gemini 2.5 Flash model enhances these interactions by better understanding conversational flows and ignoring stray sounds or voices, leading to smoother, more natural conversations with the agent. The Gemini Live API is also now integrated with Google Maps.
Google announced new ML Kit GenAI APIs that allow developers to build applications that utilize on-device AI tasks. GenAI APIs are powered by Gemini Nano. Google demonstrated re-building the Androidify app using the new GenAI APIs with just a few lines of code. The GenAI APIs can perform common tasks such as summarizing and rewriting text and generating image descriptions.
Google introduced numerous developer tools and announced major updates to existing tools this year, focused on improving developer productivity, accelerating app creation, and enhancing application capabilities. A common thread across most of these innovations is the strategic integration of AI. Here are the most notable tools:
Android Studio becomes even more powerful with enhanced Gemini AI productivity tools designed to boost developer productivity at every stage of the development lifecycle. Key enhancements include creating unit tests via natural language, updating dependencies with AI agents, and improved refactoring capabilities. Additionally, developers can now test critical user journeys across various devices with minimal code.
Stitch is an experimental AI tool that generates app designs from prompts and images. Stitch is powered by Gemini 2.5 Pro's multimodal capabilities, and, following a "design first" approach, Stitch transforms these inputs into complex UI designs and corresponding code. Users can then copy the generated markup into an IDE or Figma for further refinement.
Jules is a new asynchronous coding agent capable of handling complex tasks in large codebases, and is now in public beta. Jules directly integrates with GitHub repositories and can execute tasks autonomously. Tasks that can traditionally take developers hours to perform, such as updating a project's older Node.js version, can be performed with Jules in minutes.
Firebase Studio allows users to create an entire app using just a single AI prompt, and can even generate code by simply importing a Figma design, helping turn visual designs directly into code. And in addition to generating web and mobile frontends, Firebase Studio can also help you build full stack apps including APIs, backends, and more.
Google AI Studio empowers developers to rapidly build applications using the Gemini API. It leverages the Gemini 2.5 Pro model and allows users to generate web applications simply by providing an AI prompt. The platform includes preset prompts, such as “Build a dynamic text adventure game using Gemini and Imagen,” and features a native code editor. It automatically generates code and can self-correct detected errors, and the build process is designed to be iterative, allowing users to enhance and modify the app with additional prompts.
