AI and Crowdsourcing Tackle Crumbling Sidewalks

Project Sidewalk, a crowdsourcing platform founded in 2012, is leveraging artificial intelligence and volunteer efforts to address the widespread issue of deteriorating sidewalks across the United States. The platform allows individuals to virtually navigate streets using Google Maps Street View imagery from their own homes, cataloging problems such as broken pavement, missing curb ramps, and faded crosswalks. This crowdsourced data then fuels AI tools designed to automatically detect and classify sidewalk defects.

The initiative aims to solve the critical problem that many cities lack comprehensive data on their sidewalk conditions, posing significant risks to pedestrians, particularly those with disabilities. While the initial goal was to map all sidewalks globally, the focus has evolved to enabling volunteers to contribute data for neighborhoods throughout the U.S. Jon Froehlich, a co-founder of Project Sidewalk and a computer science professor at the University of Washington, emphasizes that while AI is advancing rapidly, it is not yet capable of acting as an expert evaluator or fully understanding lived experiences.

Instead, Project Sidewalk has become a vital tool for community engagement and local political advocacy. For example, starting in 2021, Girl Scouts and community members in Oradell, New Jersey, mapped approximately 36 miles of sidewalks, identifying over 11,000 potential issues. They subsequently presented their findings to the city council, advocating for the installation of missing curb ramps to enhance accessibility. This approach demonstrates how citizen science, combined with AI, can drive tangible improvements in urban infrastructure and accessibility.