AI Mode in Search

• With more advanced reasoning and multimodality, and the ability to go deeper through follow-up questions and helpful links to the web.
• For more thorough responses, they will be adding deep research capabilities into AI Mode in Labs."
• Live capabilities from Project Astra are coming to AI Mode in Labs. With Search Live, coming this summer, we can talk back-and-forth with Search about what we see in real-time, using our cameras.
• Agentic capabilities from Project Mariner to AI Mode in Labs, starting with event tickets, restaurant reservations and local appointments.
• AI mode will analyse complex datasets and greate graphics.
• Uses Gemini 2.5.

Gemini Features

• Interactive Quiz can be used to have a study partner who practices quizzes with students.
• Gemini Live (it's camera video conversation tool) will be able to add items to Maps, Calendar, Tasks and Keep.
• Gemini Live will have screen sharing options.
• Canvas can now create infographics, quizzes, audio overviews in 45 languages, powerful coding that can be used to prototype quickly by translating complex ideas into code.
• Upload PDFs and images into Deep Research.
• Link personal documents from Drive and Gmail into Deep Research.
• Agent Mode, an experimental feature that lets Gemini do things on your behalf (using automation). (This feature is said to be only for AI Ultra subscribers)
• Gemini in Chrome (for Google Operating Systems) will roll out soon.
• 2.5 Flash and 2.5 Pro Preview are part of the AI Plus subscription now.
• 2.5 Pro now uses Deep Think, an experimental reasoning mode best for math and coding.

Generative AI

• Veo 3 was announced and integrated into Gemini, for Ultra subscribers only.
• Veo 2 has increased capabilities now including camera controls and adding/removing objects.
• Imagen 4 is their new image generating model, available in the Gemini app, Whisk and enterprise Vertex AI.
• Flow: Google's own filmmaking model that uses their best models to create "cinematic films".
• Lyria 2: AI music generation that allows users to interactively create, control and "perform" music in real time.
• SynthID Detector, a verification portal to help people identify AI-generated content (that is watermarked with SynthID). SynthID has watermarked 10 billion pieces of content since its launch.

AI Assistance

• Project Astra: Their research prototype that explores the capabilities of a universal AI assistant now includes more natural voice output, improved memory and computer control.
• Google has partnered with the visual interpretation service Aira to build prototypes for the blind and low-vision communities.
• Astra is prototyping a new conversational tutor that can help with homework.
• Android XR - a new immersive headset - which will later include Gemini for glasses that have Android XR.

Other Announcements

Communication Technology

• Google Beam Introduced: The 3D video technology project, formerly known as Project Starline, has been renamed Google Beam.
• Industry Partnerships: Google is collaborating with Zoom, HP, Diversified, and AVI-SPL to bring Google Beam to the market for select business customers later this year.
• Speech Translation in Meet: Google Meet now has a speech translation feature available that operates in near real-time and is designed to preserve the original speaker's tone and vocal quality.

Developer Tools and Platforms

Google reported a significant increase in developer activity, alongside a broad range of new and updated tools.

• New AI Models:
  • Gemma 3n: A new open multimodal model engineered to run on personal devices like phones and laptops.
  • SignGemma: An upcoming open model focused on translating sign language into text.
  • MedGemma: A new open model for medical text and image comprehension, available now for developers.
• Coding and Development Agents:
  • Jules: An AI agent designed to assist with tasks in GitHub repositories is now in open beta.
  • Gemini Code Assist: Now generally available, this tool assists with coding tasks, including web app creation and code transformation.
  • Android Studio: Is receiving two new features: "Journeys" for testing using natural language and a "Version Upgrade Agent" to automate dependency updates.
  • Colab: Is being updated to an "agentic experience," allowing it to execute tasks from plain-language instructions.
• Tools and Platform Updates:
  • Stitch: A new tool that generates user interface (UI) designs and corresponding front-end code from text or image prompts.
  • Firebase: New features have been released, including Firebase Studio and Firebase AI Logic, to simplify the integration of AI into applications.
  • Wear OS 6: The Developer Preview is now available.
  • Google AI Edge Portal: A new solution for testing on-device machine learning is now in private preview.
  • Partnership: A new Google Cloud and NVIDIA developer community has been established.

AI Enhancements for Productivity

• Gmail Smart Replies: Later this year, Gmail will roll out personalized smart replies that use a user's past emails and files to generate responses that match their typical writing style.
• Google Vids: The AI video creation tool is now available to Google AI Pro and Ultra users.
• NotebookLM: The app is now on the Play Store and App Store. It is also gaining adjustable-length audio summaries, with video-based summaries coming soon.
• Sparkify Experiment: A new feature in Google Labs, Sparkify can turn a user's questions into short animated videos. A waitlist is open for those interested in trying it.