From Sundar Pichai's '92 Mentions Of AI' To Gemini, Google I/O Was All About Artificial Intelligence

Google I/O 2025 will be remembered for many different reasons, ranging from tech announcements to strategic partnerships, but none more than Alphabet Inc. CEO Sundar Pichai's "92 mentions of AI" in his keynote. Incredibly, this was only dwarfed by the mentions of Gemini. 

Simply put, Google I/O 2025 will be remembered for all things AI: whether it was AI's integration into Google products, or strides made in AI models, or updates to Gemini.

Here's a recap of all the AI news, announcements, and launches from Google I/O 2025.

• The AI Mode in Search will allow users to ask lengthier and more complicated questions, with advanced reasoning capabilities. 

• AI Mode in Labs will have deep research capabilities, Project Astra's live capabilities, and agentic capabilities from Project Mariner.

• New AI Mode shopping experience combines Google's Shopping Graph with AI capabilities to help users browse for ideas, compare options, and choose.

• Gemini 2.5 will be available in Search for AI Mode and AI Overviews in the US.

• With the new interactive quiz feature, users can ask Gemini to "create a practice quiz on…" and Gemini will generate questions.

• Camera and screen sharing capabilities for Gemini Live will be accessible to Gemini app users on iOS.

• Users can upload PDFs and images to Deep Research for creating research reports.

• Agent Mode will allow Google AI Ultra subscribers to describe their end goal and Gemini will get things done on their behalf. 

• LearnLM integration directly into Gemini 2.5.

• New preview version of Gemini 2.5 Flash with improved coding and reasoning performance. 2.5 Flash is now available to all in the Gemini app.

• Deep Think capabilities in Gemini 2.5 Pro for complex math and coding tasks.

• Veo 3 generates video with audio, and Veo 2 model gets new camera controls and object add and remove features.

• Imagen 4, the latest picture generation model, offers clarity in details like skin, fur and intricate textures.

• Flow, a new AI filmmaking tool, will allow users to expand a brief clip into a longer scene and make cinematic films.

• SynthID Detector will make AI-generated content detection easier using watermarks.

• First Android XR device to arrive later this year, Samsung's Project Moohan, offering immersive experiences on an infinite screen.

• Gemini-powered glasses with Android XR will help in messaging, making appointments, asking for directions, taking photos, etc.

• Android XR glasses can help overcome language barriers, which Google demoed with a live language translation.

• Updates to Project Astra include more natural voice output with native audio, improved memory and computer control. Also, currently in prototype, a conversational tutor will help with homework. 

• Google Beam, a new platform for video communications, will use cameras and AI to combine 2D video feeds into a realistic 3D experience.

• Speech translation is now available on Google Meet, and can mimic the speaker's tone, voice, and even facial expressions in almost real time.

• Smart Replies in Gmail will capture the user's tone, manner and preferred word selections to generate a reply.

• NotebookLM app, available on Play Store and App Store, will help users take Audio Overviews on the go, and it will soon have Video Overviews.