In This Story
Google CEO Sundar Pichai on Tuesday announced an AI tool that could very well quell the pleas of every young person who’s watched a parent scroll through photos on their smartphone for long minutes in search of “that one time when…”
The new Gemini feature, Ask Photos, allows users to task their AI assistant with finding photos. Prompts might include “Show me the best photo from each national park I’ve visited,” or “What themes have we had for Lena’s birthday parties?” In response, the chatbot will chat back with relevant photos, “saving you from all that scrolling,” the company said in its announcement during its annual Google I/O developer conference.
Pichai demonstrated how customers can also use Ask Photos to retrieve specific information from within their photos. For example, he displayed, a user can ask the chabot to remind them of their license plate number; the AI will analyze pictures of their car within their camera roll to generate a response. “It knows the cars that appear often, it triangulates which one is yours, and just tells you the license plate number,” Pichai said.
So how secure is an AI that sifts through our photos? Google addressed “[W]hile Ask Photos is experimental and will not get everything right, we employ layers of safeguards and AI models to help ensure responses are safe and appropriate,” Google wrote in its online announcement of the product.
But it also added that the tool will remember conversations, using corrections or extra information you provide to train itself for the future.
Google plans to roll out the feature this summer “with more capabilities to come,” said Pichai.
The Google chief also made several other announcements of new Google AI features, including the official rollout of AI overviews in Google Search, a new image generator Imagen 3, and a new AI chip called Trillium. It also demoed an AI assistant that can chat through a user’s phone or smart glasses.