Google announced a host of new artificial intelligence features coming to its products and services at the company’s annual I/O developer conference yesterday (May 10), where CEO Sundar Pichai mentioned the term “A.I.” 27 times during his 15-minute keynote, per Observer’s count. (Other executives mentioned “A.I.” about 100 additional times over the two-hour event.)
At yesterday’s event, Google officially unveiled PaLM 2, its new large language model (LLM) that has improved capabilities in reasoning, coding and translation compared with the previous LaMDA language model.
PaLM 2 powers more than 25 Google services, including the Bard chatbot, Google said. Here are five biggest A.I. announcements from yesterday’s presentation.
Google Search enhanced with A.I. snapshots
Google is introducing a new feature called AI snapshots to its flagship search engine. Google Search enhanced with AI snapshots will show A.I.-generated answers at the top of search results for some queries to provide relevant context. Users can further narrow down information in the snapshot with follow-up questions.
A.I. snapshots will be available in the coming week through a product called “Search Generative Experience” (SGE). Users need to first sign up for a testing program called Search Labs, which opened yesterday.
More features are coming to Bard
Google said yesterday its updated Bard chatbot is now available to everyone. The new Bard, powered by Google’s PaLM 2 language model, is also adding new features, including support for Japanese and Korean languages and a quick way to export generated text to Google Docs and Gmail.
Google also announced plans to introduce image generation functionalities powered by Adobe’s A.I. image generator, Firefly.
Android’s A.I.-powered customization options
A series of A.I. features coming to Android, Google’s mobile operating system, will allow users to customize their smartphones more easily than ever.
For example, Android phone users will soon be able to reply to text messages with A.I.-generated responses through a tool called Magic Compose. They can also create unique screen wallpaper using Google’s text-to-image diffusion model.
Next month, Google will launch an exclusive “cinematic wallpaper” feature for Pixel phone users. It will let them customize wallpaper by adding depth to existing photos.
Google Photos introduces Magic Editor
Google Photos, Google’s photo sharing and storage service, will introduce new editing functions that allow users to perform advanced Photoshop-like actions all on their smartphone. For instance, users can move around a person or object in a photo, erase random people in the background, and enhance environment colors.
The feature will come to select Pixel users first later this year.
Duet AI for Google Docs, Sheets, Slides, Meet and Gmail
Google has been working on a host of A.I. features for its Workspace suite of apps, like Docs, Sheets, Slides, Meet and Gmail, for several months to compete with Microsoft’s similar effort with its Office software.
At yesterday’s event, Google announced the official branding of this effort: Duet AI. Duet AI will include tools like writing assistance in Docs and Gmail, an image generator for Slides, and automatic meeting summaries for Meet.
To access the new features, users need to sign up to Workspace Labs and join a waitlist.