Google is fully committed to advancing artificial intelligence, making a clear statement during its keynote at the I/O developer conference on Tuesday, where "AI" was mentioned over 120 times. That's a significant emphasis!
However, not all AI announcements from Google were groundbreaking; some were more incremental or recycled ideas. To help you navigate through the important updates, we've compiled the most noteworthy AI products and features introduced at Google I/O 2024.
Generative AI in Search
Google is set to leverage generative AI to enhance the organization of entire Google Search results pages.
What will these AI-organized pages look like? It will vary based on the search query. Google mentioned that users might see AI-generated summaries of reviews, discussions from social media platforms like Reddit, and curated lists of suggestions. Initially, AI-enhanced results will primarily appear when users seek inspiration—such as during trip planning—and will soon extend to searches for dining, recipes, movies, books, hotels, e-commerce, and more.
Project Astra and Gemini Live
Google is enhancing its AI-powered chatbot, Gemini, to improve its real-world understanding. This includes a new feature called Gemini Live, enabling users to engage in "in-depth" voice chats on their smartphones. Users can interrupt Gemini while it speaks to ask questions, and the chatbot will adapt to their speech in real time. Additionally, Gemini can view and respond to users’ surroundings using their smartphone cameras, facilitating questions about nearby objects or locations.
Gemini Live is scheduled for release later this year and will leverage technologies from Project Astra, which aims to create AI applications and "agents" for real-time, multimodal interactions.
Google Veo
With the introduction of Veo, Google is competing with OpenAI’s Sora. Veo is an AI model capable of generating 1080p video clips around one minute long based on text prompts.
This model can emulate a variety of cinematic styles, produce landscape shots, and perform complex edits on already generated footage. It understands various camera actions and visual effects through descriptive prompts. Veo also has insights into physics—like fluid dynamics and gravity—enhancing the realism of its creations. Additionally, it supports masked editing for specific video areas and can produce longer videos from a series of related prompts.
Ask Photos
Google Photos is receiving a new AI feature called Ask Photos, powered by the Gemini family of generative AI models, set to roll out later this summer.
This functionality will enable users to search their Google Photos collection using natural language queries based on the content and metadata of their images. For example, instead of searching for a specific landmark, users can query for “the best photo from each National Park I visited.” Gemini will assess factors like lighting and clarity to identify the "best" photos while incorporating geolocation data.
Gemini in Gmail
Gmail will become more intuitive with the integration of Gemini, which will assist users in searching, summarizing, and drafting emails, as well as managing complex tasks like processing returns.
In one demo at I/O, Google illustrated how a parent could use Gemini to summarize recent school emails. In addition to summarizing email content, Gemini will analyze attachments—like PDFs—and generate concise summaries, highlighting key points. Furthermore, Gemini will help organize receipts from emails and extract information for use in Google Drive or spreadsheets, streamlining workflows—ideal for those who frequently track expenses.
Detecting Scams during Calls
Google introduced an AI-powered feature to alert users about potential scams during phone calls.
This feature will be incorporated in an upcoming Android version, utilizing Gemini Nano—the smallest model in Google’s generative AI lineup—which operates entirely on-device to monitor for real-time conversation patterns typically associated with scams. There is currently no specific release date, but the feature will be opt-in, which is a positive privacy measure.
AI for Accessibility
Google is enhancing its TalkBack accessibility feature with generative AI capabilities.
Soon, TalkBack will use Gemini Nano to generate auditory descriptions of objects for users with low vision or blindness. For example, it could describe clothing like this: “A close-up of a black and white gingham dress featuring a collar and long sleeves, tied elegantly at the waist with a large bow.” According to Google, TalkBack users encounter about 90 unlabeled images daily, and the integration of Nano will significantly enhance accessibility.
We’re launching an AI newsletter! Sign up here to receive the latest updates directly in your inbox on June 5.
Ranking
