Hivemapper’s rewarding approach to map accuracy combines human insights with AI efficiency.
Maps are essential in our daily lives, whether we’re navigating new cities or searching for the nearest coffee shop. Traditionally, maps were created by teams of cartographers and surveyors who would collect data and draw maps by hand. These physical maps eventually became the digital maps we know today.
Although maps have been digitized, OpenStreetMap and Waze still rely on thousands of humans to painstakingly edit and annotate maps in the manner that you see below. Hivemapper believes in a more efficient approach that combines AI and humans.
Caption: This GIF shows how an OpenStreetMap user edits the map. This tedious method will soon be replaced by AI.
Unlike Big Tech companies that take user data without compensating map editors, the Hivemapper Network rewards contributors with HONEY tokens for collecting street-level imagery and refreshing it. Our collection device, the Hivemapper Dashcam
, is 1,000 times cheaper than a Google Street View car.
Hivemapper’s current growth rate of coverage is roughly 8x faster than the 133,000 kilometers per month that Google averaged in the first five years after the launch of Street View in May 2007. As of Sept. 2023, Hivemapper contributors have collected more than 50 million total km in 8 months
, 5 million being unique roads.
But continuously updated imagery is only the raw ingredient of our map. The real magic is rooted in how we convert this imagery into an accurate and detailed global map — tapping into a $300 billion market
that’s currently dominated by Google Maps.
Object Detection and Classification: Help the Map AI Improve
Our Map AI processes the imagery drivers collect with our purpose-built dashcams and extracts map data: traffic lights, speed limits, turn restrictions, highway exit information, etc.
Map AI Trainer games help our Map AI produce quality map data. As contributors play these games, they train the Map AI to learn things like what a speed limit of 25 mph looks like, what a no U-turn sign looks like, and even what a bad dashcam mount looks like. This process requires many contributors to review before reaching consensus on any given object.
Play AI Trainer Games & Earn HONEY
If you don’t have a car or can’t afford a Hivemapper Dashcam, you can still contribute to the Hivemapper Network and the future of mapmaking by playing AI Trainer games, which are designed to be easy to learn. Anyone can play using a computer or mobile device.
AI Trainer games are an innovative solution for building a map that is accurate, up-to-date, and accessible to everyone. By participating in these games, you can help improve the quality of our map data while earning HONEY. Players are rewarded with HONEY for quality submissions. If someone is just clicking and submitting nonsense reviews, they will quickly be blocked.
Start playing AI Trainer games here.
Why AI Trainer Games Are Important
Hivemapper's rapid pace of collection has caught the attention of prospective customers, who are now interested in verifying the quality and precision of the extracted objects. In order to meet the demands of these customers, the Hivemapper Network is expanding to accommodate millions of Map Editing and AI Trainer tasks every week, ensuring the availability of an up-to-date and precise database of road objects.
By playing AI Trainer games, you can help improve the accuracy and quality of Hivemapper’s map data while earning HONEY.
Join the Community
If you have any questions about Hivemapper’s AI Trainer games, reach out to us on Discord
or email us at firstname.lastname@example.org