On Tuesday, the company announced in Google I/O that Google was adding its Gemini AI assistant to Chrome.
According to a blog post by Google Labs and Gemini VP Josh Woodord, Gemini will “be able to clarify complex information on any web page or summarize the information.” Google imagines that Gemini will later “work on numerous tabs and visit the websites by you.”
I saw a demo during a briefing before Tuesday’s announcement. In Chrome, you will see a slightly bright icon in the right corner. Click on it and open a Gemini chat boot window – this is a floating UI that you can move and change size. From there, you can ask questions about the website.
In the demo, Charman de Silva, director of the Product Management in the Chrome team, opened a page for a sleeping bag in Rei and clicked on a recommended Gemini prompt for a list of key features of the bag. Gemini read the whole page and entered an immediate summary of the bag. De Silva then asked if the sleeping bag was a good option for camping in the mine, and in Chrome, Gemini responded by pulling information from the Rei page and the web.
Subsequently, De Silva went to a shopping page on another retailer website for a different sleeping bag and asked Gemin to compare two sleeping bags. Gemini did this and included a comparison table.
When you go from the tab to the tab, you will be able to continue the conversation with Gemini initially. But “at the end of the year,” Gemini in Chrome will allow you to choose multiple tabs at once and ask a question about them all.
De Silva also showed a feature demo that will be available in the future: the use of Gemini to visit the website. In the demo, De Silva pulled Gemini directly in Chrome to help visit the site. De Silva asked Gemini to scroll the ingredients, and AI zipped this part of the page. He also responded when De Silva sought help to convert the required sugar quantity into grams.
In Google’s selected demo, Gemini in Chrome looks as if it can sometimes be useful, especially with comparison tables or momentarily conversion. Instead I just read the website or do my research instead of reading Gemini’s AI summary, especially since the AI can deceive the false information.
Gemini is launching on Wednesday in Chrome. It will initially release Windows and Macos in early access to 18 or older users who use English as their own language. It will be available to people who subscribe to Google’s AI Pro and Ultra Subscripts or Chrome’s Son, Canary, and Dev Channels.
“This is an area we will think about,” Tabriz says, “It is a matter of bringing Gemini, but now, the company is” focusing on the desktop. “
Correction, May 20: Gemini in Chrome can continue the conversation when you go from the tab to the tab. As we have initially reported, it just doesn’t work in two tabs.
