Google I/O 2024 was packed with exciting updates this year (more than 100, according to Google themselves). Unsurprisingly, most announcements were linked to Google’s evolving software (such as the Workspace suite), generative AI, and the Gemini large language models.

Even at a glance, it seems clear that Google is focused on strengthening its position in the AI landscape, thanks to increasing pressure from Microsoft, Meta, and even Zoom. However, Google’s I/O event certainly gave us a sneak peek into how the tech giant is envisioning the future of work.

Here’s a comprehensive list of the most significant reveals, updates, and announcements from Google I/O, specifically tailored to the interests of business leaders.

Google I/O 2024 Gemini Updates: The Gemini Revolution

For the last few months, Google has focused heavily on promoting Gemini, its range of large language models and apps designed to compete with solutions like Copilot. During Google I/O 2024, Google drew attention to numerous Gemini evolutions.

Most of the Gemini announcements, however, revolved around enhancements to the Gemini app and how Gemini will be implemented into existing Google tools.

Updates to the Gemini App

Google’s Gemini app is essentially an upgraded version of the Google Assistant and an evolution of the Bard generative AI chatbot. At Google I/O 2024, Google shared an update to the model powering the app (Gemini Pro). Gemini 1.5 Pro, now available to Gemini Advanced subscribers, can analyze documents, videos, audio recordings, and codebases longer than before.

The solution has a 1 million token context window (the largest of any commercially available chatbot worldwide). Google also introduced features like:

  • The ability to upload files from your device or Google Drive to Gemini Advanced
  • Data analysis via Gemini for uploaded files and spreadsheets
  • An advanced planning feature (powered by Gemini) for travelers
  • Gemini in Google Messages, so you can chat to the bot while messaging friends

Plus, we now have Gemini Live for Advanced subscribers, which uses speech technology to enable voice conversations with the bot. Live allows users to choose from 10 human-sounding voices. Gemini Advanced subscribers will soon be able to create “Gems”, customized versions of Gemini, similar to the bots you can make with Microsoft Copilot Studio.

Google Gemini in Project Astra

While Google announced various updates to it’s AI strategy at Google I/O 2024, one of the most exciting was the development of Project Astra. This ties the Gemini model into cameras on smartphones, enabling it to interpret and understand the world around it.

Google says Astra is its “vision for the future of AI assistants.” It can identify speakers, read code, explain images, and more. It can also be added to smart glasses, potentially allowing users to interact with content in a brand-new way in the years ahead.

There’s no timeline for when this new solution will be generally available, but Google is working with developers in Singapore and Paris for a 6-month trial period.

Google Gemini in Google Search

Most of us rely on Google search every day for research, marketing, and countless other purposes. Even before Google I/O 2024, it seemed likely that Google would introduce more of its AI technology into the search engine this year.

During the event, the company confirmed it was using a new Gemini model (customized for search) to bring multi-step reasoning, planning, and multimodality into the engine. Alongside that, AI overviews in search are now rolling out in the US, and multi-step reasoning options are coming to the AI overviews in Search Labs, for US customers.

Google also says users will soon be able to adjust their AI overview to simplify topics or break them down in more detail. Plus, you’ll be able to ask questions with videos and access AI-organized results pages specifically designed to drive you toward the right content faster.

Google I/O 2024 Workspace Updates

For users of Google Workspace, the company’s productivity suite, Google I/O 2024 outlined various changes to the apps teams use daily. Notably, Gemini, or versions of it, have been available in various parts of Google Workspace for a while now.

However, in the months ahead, we’ll see more major updates to the AI functionality baked into Google Workspace apps, such as Meet, Gmail, and Photos. For instance, Gemini 1.5 Pro will now be available in a side panel for Gmail, Drive, Docs, Slides, and Sheets through Workspace Labs.

This feature, which will allow you to interact with the Gemini chatbot as you work (similar to Microsoft 365 Copilot), will also be rolling out to Google One AI Premium and Gemini for Workspace customers. In the future, users will be able to:

  • Summarize Gmail messages with action items.
  • Generate responses with contextual smart reply and Gmail Q&A
  • Write content in Gmail and Docs with additional languages (Spanish and Portuguese)
  • Organize messages and email attachments in Drive with Gemini
  • Use the “Ask Photos” app to find specific images in your cloud storage system
  • Create highlight galleries with personalized captions using generative AI

Notably, Google also introduced its “AI Teammate” solution at Google I/O 2024, although we’re not certain when this feature will be available. What we do know is that the feature will essentially give companies access to an AI employee, with its own Google Workspace account, that can complete various tasks.

The AI Teammate can appear alongside other employees in chat groups, emails, and documents, and collaborate with staff members.

Other Major Generative AI Updates from Google I/O 2024

Although much of Google’s event looked at the growing potential of Gemini, the company also drew attention to other experiments it’s running in the AI landscape. The organization announced the arrival of Trillium, its sixth-generation custom AI accelerator, for AI development.

It also introduced Grounding with Google Search, now generally available on Vertex AI. Other major generative AI updates and experiments included:

  • Imagen 3: Imagen 3 is Google’s highest-quality model for image generation, capable of understanding natural language, intent, and longer prompts. It can even render text (unlike most AI models). This solution has been rolled out to Trusted Testers of ImageFX, and you can sign up to the waitlist here or wait until it arrives on Vertex AI this summer.
  • Google Veo: Google’s new Veo solution is a video generation model capable of creating 1080p video based on natural prompts. Some of the capabilities of the solution will also be introduced to YouTube Shorts and other Google products going forward.
  • Music AI Sandbox and MusicFX: The Music AI sandbox is a suite of tools that allows users to create instrumental audio from scratch, transfer styles between different trackers, and more. MusicFX, Google’s audio creation tool, has been updated with a new DJ mode.
  • VideoFX: VideoFX is a new experimental tool from Google that uses the Veo model to turn ideas into a video clip. It includes a storyboard mode, which allows users to edit content scene by scene and introduce their own audio.
  • ImageFX: Google’s ImageFX will now have more editorial controls and use Imagen 3 to unlock additional photorealism during rendering.

Google I/O 2024 Android Enhancements: Security and More

Surprisingly, Google didn’t announce any new smartphones at this year’s event. However, they did provide many insights into how the Android operating system is evolving. First, mobile employees using Pixel devices will now have an updated version of Gemini Nano with multimodal capabilities. This means your phone will be able to understand images, spoken language, and sounds.

Google’s mobile devices will also be more accessible with the “Talkback” feature. This uses AI to describe images to users in detail. Perhaps most importantly, Google is also introducing features to Android OS to help deal with spam and scam calls. A new opt-in feature uses Gemini Nano to listen for conversation patterns typically associated with scams in real time.

There’s even a new “Theft Detection” lock feature, which identifies motion commonly associated with theft to lock your phone. Other noteworthy Android features include:

  • “Ask this PDF” for Gemini Advanced users to help users draw information from documents without scrolling through multiple pages.
  • The ability to create and drop AI-generated images into Google Messages and Gmail and ask for information about videos you view.
  • Updates to Circle Search, which will allow users to solve more complex problems with AI involving symbolic formulas and graphs.
  • Private Space- a feature that allows users to create a secure environment where they can keep certain apps protected with an extra layer of authentication.
  • Augmented reality features in Google Maps to enhance navigation in the real world.
  • Updates to Wear OS 5, reducing battery life consumption and enhancing data support.

Google’s Updates for Developers

Finally, Google also shared a range of updates for developers leveraging its technology ecosystem. For instance, the company previewed the next version of Gemma, its family of lightweight open AI models. Gemma 2 is built on a new architecture with a larger 27 billion parameter instance.

There’s also the new PaliGemma solution, a vision-language model optimized for image captioning and visual Q&A experiences. On top of that, Gemini models are now available to developers in Android Studio (Gemini 1.5 Pro), IDX, Firebase, Cloud, Collab, and IntelliJ.

A few additional notable announcements include:

  • Parallel video and calling frame extraction with Gemini’s API
  • Context caching via the Gemini API solution
  • Gemini Nano built into the Chrome desktop client
  • Project IDX (now generally available) for creating full-stack multiplatform apps
  • Firebase Data Connect for developers using SQL with Firebase
  • Kotlin Multiplatform support for Android developers sharing apps across platforms

What’s Next for Google?

This year’s Google I/O 2024 was packed with interesting updates and releases. For the most part, it seems like Google is focusing heavily on maintaining its edge in the AI space.  The event was packed with new developer tools, Gemini models, and generative AI experiences.

The good news is Google is focused on preserving privacy, ethics, and security with its innovations. The company announced this year that it’s enhancing its “red teaming” practice to evaluate models for weaknesses.

The company also said it would be partnering with educational leaders on its LearnLM solution, specifically tuned to the needs of learning institutions. The family of LearnLM models, based on Gemini, are custom-made to protect and preserve sensitive data in the educational landscape.

If you want to learn more about all of the announcements from Google I/O, you can find Google’s own blogs, videos, and demos online, for deeper dives into each new feature. Alternative, stay tuned for more reviews, insights, and articles here on UC Today.



from UC Today https://ift.tt/GRsrboM