What is Google Gemini Ask Photos? At Google I/O the company revealed a new way to leverage your photo library. Here’s how it’ll work.
One of the first key announcements in Google’s AI-dominated I/O keynote event was ‘Ask Photos’. The tool, which is rolling out to the Photos app in the coming months, will enable users to ask the AI assistant to search their photo library to surface complex results – far beyond the usual keyword searching.
Google said during the keynote that over 6 billion photos are uploaded to Google Photos every day and Gemini can leverage those to cut down on scrolling and deliver results across your entire library, with extremely deep context.
For example, via a new tab within the Google Photos app, the user could ask: “Show me the best photo from each national park I’ve visited” and Gemini would yield precisely that, without the need to scroll through the entire library, or search for specific locations.
The response is incredibly detailed. For example, it might tell users: “Here’s a great photo from each of the 22 national parks you’ve visited around the world, including: Experiencing the Amazon at Yasuni, Peaks and lakes of the UK, Each of Utah’s ‘Mighty 5’ national parks.”
During the keynote address, Alphabet CEO Sundar Pichai asked Gemini to bring up photos of a young girl called Lucia through the lens of her swimming journey. From her first dip and swimming lessons, to her first swimming certificate. It would be surfaced in an album, cutting plenty of time searching via keywords.
Pichai also spoke of the ability to ask Gemini “what’s my license plate number again?” Gemini will search the library for pictures of cars, identify what it deems to be your current vehicle and show you an image of the license plate.
Google says it’ll help users plan original birthday parties for their keys by asking Gemini of the themes they’ve already used. It may respond: “Here are the themes from Lena’s last 4 birthdays: 3rd: A princess celebration, 4th: Under-the-sea with mermaids, 5th & 6th: Two magical unicorn parties.“
Google says the key is the ability to understand the question and complex natural language terms in order to conduct a more sophisticated search of the library. Keywords, places, people and dates are picked up, but also concepts like “themed birthday parties”
The second element is the crafting of the response, which comes through studying the results of the search, assigning relevance (sometimes by reading text within the images) and choosing which photos are surfaced in the results.
Google says the final element is to ensure responses are safe and appropriate. It’ll also respond to your corrections and additional context and remember for next time.
You’ll be able to try Google Gemini Ask Photos this summer.
