The technology is more sophisticated than Snapchat's Bitmoji feature. Google's team used neural networks (code based on the way our real-life brains work) and machine-learning algorithms (programs that adjust themselves based on new information) to figure out how to make the best emoji replica of each user's face.
The cartoonish designs let Google Allo make the images distinctive enough to look like specific people but not so realistic that they're creepy. The resulting emoji are "less about reproducing reality and more about breaking the rules of representation," Allo's expressions creative director Jennifer Daniel wrote in a blog post.
Artists drew a bunch of different features that the app can match to each selfie. For example, it'll give your emoji the hairstyle that looks most like yours. The technology also lets Google Allo create an image to represent you that's not influenced by temporary factors like what kind of lighting you're in.
Representing diversity was also important to the team behind the feature. "Illustration by its very nature can be subjective. Aesthetics are defined by race, culture, and class, which can lead to creating zones of exclusion without consciously trying," Daniels wrote. "As such, we strove to create a space for a range of race, age, masculinity, femininity, and/or androgyny. Our teams continue to evaluate the research results to help prevent against incorporating biases while training the system."
If Allo's downloaded on your Android, you should be able to create customized emoji already. The feature also plans to launch on Allo for iOS soon. In the meantime, iPhone users will have to occupy themselves with Bitmoji.