If you're the kind of person who spends a chunk of their day looking at animated GIFs, a pair of students from MIT want to use your skills to create a new universal language. Kevin Hu and Travis Rich have come up with GifGif, an MIT Lab project designed to extract the very essence of various human emotions captured in GIFs by running them by actual humans. The project, which was spotted by FastCo Design, pits two GIFs head-to-head, and asks which one is the better representation of a human emotion. Choosing the better one (or neither) loads up two more GIFs, and you can cycle through various emotions to your heart's content.

The project currently indexes 4,653 GIFs, and is adding new ones as certain GIFs are retired after being "well-characterized" by users. It's pulling from Betaworks' Giphy GIF repository, which is already categorized and searchable. Hu and Rich say they eventually plan to open up the service to user uploads, as well as for developers to integrate the search results in their own software, something that could let people find the appropriate GIF to convey meaning in all sorts of scenarios — even if they didn't speak the same language. In the meantime, the pair is offering those who aren't simply seeking quizzes a search tool to find specific GIFs based on a combination of 14 different emotions.