Some of the most innovative software applications and services from the past decade have been powered by collective intelligence. Google, Wikipedia, and Waze, to name just a few, work by intelligently coordinating the collective behaviors of thousands, if not millions, of people on the Internet.
Koko applies collective intelligence to the domain of well-being, using an approach that was originally conceived at the MIT Media Lab. In a randomized controlled trial, a web-based version of Koko outperformed an existing intervention on a host of psychological outcome measures.
In recent years, crowdsourced data has helped unlock new forms of artificial intelligence. Crowdsourced traffic annotations from Waze, for example, have helped power new map APIs. Koko follows a similar approach and uses its crowdsourced data - a massive repository of human kindness - to unlock new forms of artificial, emotional intelligence. We use the output of our peer-to-peer system to give machines the ability to provide nuanced, empathetic support.
Koko is designed and built in New York City by Kareem Kouddous, Rob Morris, Fraser Kelton, James Simonian with support from Rachel Nash. We are funded by USV and Omidyar Network and we are brought to life by our vibrant community, which includes thousands of people from over 150 countries across the world.