Google I/O 2024: Gemini AI at the core of user experience ✨

Cliquez ici pour lire en français

The Google I/O 2024 conference was marked by a number of major announcements concerning Gemini artificial intelligence. In fact, the word « AI » was mentioned 120 times during the conference. Developed in-house by Google, this powerful generative AI is at the heart of the new strategy to rethink the user experience across all products.

Gemini AI and Google search 🔍

The star of the show was undoubtedly Gemini 1.5, available in different versions. Gemini 1.5 Pro can process huge data sets of up to 1 million tokens, paving the way for ultra-detailed analyses. A lightened version, Gemini 1.5 Flash, was also presented for light tasks requiring very low latency. Finally, Gemini Advanced, a premium model with advanced capabilities, was launched as a preview in France.

Gemini AI will profoundly transform the Google search experience. With « AI Overviews », results will offer complete, instant answers to complex questions, compiling different perspectives and sources. The engine will also be able to break down a query into sub-questions to provide a detailed answer. Search also integrates new intelligent planning and video comprehension capabilities.

More intelligent photo search 🖼️

Google Photos is also getting a whole lot smarter, thanks to Gemini. The « Ask Photos » function will enable contextual searches in natural language. It also analyzes images and metadata to identify the best photos matching the query. With Gemini, Google Photos gets smarter:

• Photo sorting: Find the best photos of visits to national parks by analyzing light, blur, etc.
• Contextual questions: Ask for contextual information such as the themes of previous birthday parties.
• Image text analysis: Interpret text in photos to answer specific queries, such as license plate numbers. This feature will first be available in the U.S. in English this summer.

Gemini will make you more productive🚀

Gemini AI also joins the Google Workspace suite. Its integration of Gemini will improve your productivity with automatic e-mail summaries and intelligent replies. You’ll also be able to get meeting summaries in Meet.

Google has also unveiled NotebookLM, a multimodal voice assistant based on Gemini 1.5 Pro. It features natural sound and fluid interaction, just like OpenAI (which you can discover by clicking here). Designed with particular attention to confidentiality, it can :

• Summarize key facts and ideas from a text.
• Generate links between sources.
• Customize your workspace by integrating Google Docs.

The Astra project to improve human-machine interaction🤖

Google is making progress in the creation of AI agents for more natural and effective assistance. A demo of the Astra project was presented. The AI agent :

• Understands context, offers low latency and is fast.
• Uses phone camera and AI to find lost objects, explain how objects work, etc.
• Can include smart glasses.

Gemini Nano against scams ⚠️

Google also introduced a Gemini Nano feature to warn against scams. It works locally on the smartphone to listen in on phone conversations and warn the user of potential scams. This feature is currently in the testing phase.

AI for creativity 🖌️

In the field of content creation, Google has lifted the veil on Imagen 3 for photorealistic image generation, Music AI Sandbox for AI-assisted music creation, and Veo for high-quality video generation. An AI watermark called SynthID will make it easy to identify AI-generated content.

Learn better with AI 👩🏾‍🏫

Google hasn’t forgotten education either, with LearnLM, a family of AI models optimized for learning and aimed at making Google products more engaging for students and teachers.

Google I/O is also an opportunity for Google to present the new features coming to Android. We talk about them in a dedicated article available by clicking here.

Overall, this Google I/O 2024 clearly highlighted Google’s determination to place Gemini generative AI at the heart of the user experience across all its products, from search to productivity to content creation. An ambitious gamble to rethink our interactions with technology. What do you think?