How Google is counting on AI to provide a boost to its suite of everyday tools

By Courtesy of Google Google I/O 2023

This year, artificial intelligence was at the heart of the big launches and announcements at the Google I/O conference. The American giant presented its projects of integrating AI into its various tools and applications. Between personalized searches, easy photo editing and automatically written email messages, internet users may be seeing some significant changes in the near future.

This year, artificial intelligence was at the heart of the big launches and announcements at the Google I/O conference. The American giant presented its projects of integrating AI into its various tools and applications. Between personalized searches, easy photo editing and automatically written email messages, internet users may be seeing some significant changes in the near future.

A new language model

Google this week presented its brand new language model, PaLM 2, set to be efficient at a wide range of tasks and easy to use. Compared to its big rival GPT-4, PaLM 2 is particularly powerful in terms of logic and scientific and mathematical reasoning. It can also, theoretically, help developers code or fix a bug.

Specifically tailored models have already been developed, such as Med-PaLM 2 for the medical field and Sec-PaLM, optimized for cybersecurity. The advances in Google's language model will obviously help boost Bard, its conversational robot, a competitor of ChatGPT. The goal is to make Bard available in the 40 most spoken languages in the world by the end of the year.

A more conversational search engine

For the general public, Google's AI progress should become rapidly noticeable in many of the tech giant's main everyday tools. The famous search engine will make use of these new generative artificial intelligence capabilities to optimize its results, with the goal of being better able to understand a subject, with more information provided as well as various points of view on it expressed. Google takes a complex question like "what's better for a family with kids under 3 and a dog, Bryce Canyon or Arches [National Park]?" as an example. Here, generative AI will provide a comprehensive, practical answer with further steps suggested and context carried over. The conversation can then continue with other practical jumping-off points on the same topic. The AI should also help to simplify online shopping, by offering a summary of the key criteria to be taken into account as well as a list of products responding to various requests. A first experiment will be launched in the United States, with Chrome and the Google app (Android and iOS), for initial feedback.

Immersive 3D views in Maps

Maps will also benefit from greater integration of AI, through its new feature called "Immersive View." This will allow for the creation of a high-resolution representation of a place or a route, in order to allow the user to explore it virtually before actually going there. To do this, the AI processes billions of photos and aerial images and creates a dynamic 3D model that allows users to get a preview of nearly anywhere and find a multitude of practical information about the spot, such as the presence of bike paths, sidewalks or parking lots before arriving at the destination. Other information will also be provided such as real-time weather, air quality and traffic conditions. This new feature will be rolled out in the coming months for trips in Amsterdam, Berlin, Dublin, Florence, Las Vegas, London, Los Angeles, New York, Miami, Paris, Seattle, San Francisco, San Jose, Tokyo and Venice.

Smarter photo editing

Google also plans to introduce a new experience this year called Magic Editor in Photos. The idea is to be able to automatically recreate entire parts of the image that were not captured in the original shot. In principle, it should be possible to remove elements, people or objects, from the original photo, with the artificial intelligence taking care of filling in the hole or space that results. This new feature should also allow users to easily change the brightness and even the color of the sky, to "clean" a dirty garment, remove reflections in eyeglasses, etc.

Write automatic email messages in Gmail

New generative AI features will also make an appearance in Gmail. While it is already possible to send short ready-made answers or to benefit from suggestions as you type your message, Google is going to speed things up with a new feature presented as "Help me write." To explain its possibilities, Google takes the example of receiving an e-mail informing you that your flight has been cancelled, accompanied by a voucher. However, if you want a full refund instead, all you have to write a prompt for the message to be created, in this case, "ask for a full refund for this canceled flight," and Gmail will create a draft ready to be modified and sent. The text will then include various contextualized elements in its composition. This option will be launched soon, initially via an update to its Workspace business suite.

© Agence France-Presse