“Best Take is great for putting together the perfect group shot — and for having some fun,” says Pixel Camera Product Manager Isaac Reynolds. “Making a so-called ‘worst take’ with funny expressions all around is my favorite use case — you have to choose which of your ridiculous expressions is the best yourself, and it’s pretty funny to see what you can create!”

Best Take can trace its origins to a week-long summit when teams from Google Photos, Pixel and Google Research decided to tackle what they call the “group shot dilemma.” User research told them that capturing the ideal group photo was a challenge, so they brainstormed how to help, ultimately landing on the idea of an AI-powered tool that combines different facial expressions from pictures you’ve taken into one image.

Early concept testing showed that the feature appealed partly because it tapped into a common behavior: People already take multiple photos when taking group shots. “A big reason this feature resonates with people is that the composite image feels true to their memory — by drawing from pictures taken in quick succession, it’s essentially the same moment,” says Photos Product Manager Lillian Chen. “The feature balances authenticity with helpfulness to solve a real pain point.”

Best Take is driven by a combination of different AI models that analyze images, check timestamps to find photos taken in sequence, search for signals like head poses and facial expressions and finally suggest images of faces where subjects are looking at the lens and smiling. Once selected, technology adapted from AI-driven Google Photos features like Magic Eraser and Cinematic photos helps seamlessly blend the images and fix any background errors.

Beyond providing an automatic suggestion, the design prioritizes user choice as well. “Our goal is to give users control over their memories, including which alternative expression they prefer. Even the most subtle differences can matter to each person in the photo” says Research Product Manager Navin Sarma.