Elephant Butts and Poverty
The quantified self has become the quantified planet.
Last year at Google Zeitgeist, I reconnected with a friend of mine who runs a firm that manages large data sets.
We started talking about Samasource’s work doing image tagging for machine learning. So many technology innovations — from drones to self-driving cars to facial recognition on your iPhone — depend on high-quality, detailed image annotation to train algorithms to detect objects and activities. We’ve trained teams of people in rural Northern Uganda and in urban slums to carefully trace objects in images, tag items, and label them. It’s surreal to see someone who grew up doing subsistence agriculture earning a great wage by tagging pictures of Rihanna or plotting points on a human hand.
My friend came to us with a super interesting problem: how can we use real-time image data from satellites to count elephants in the wild and identify threats to conservation, such as poaching? They had been working on the Great Elephant Census, a massive undertaking by Paul Allen to count all the elephants in the wild. So we began an ambitious project to tag as many images as we could, drawing a rectangle around each and every…elephant butt.
Yes, that’s right. A team of half a dozen people in Northern Uganda are looking for elephant butts and earning a living wage while they’re at it. The posteriors of these gentle giants are their unique identifier. Our workers love this project, because it’s a chance to give back meaningfully close to home.
Most of our workers were affected by the Northern Ugandan Civil War for much of their lives, and they were deprived of the chance to contribute meaningfully to their economy. Doing work like this, which is so vital to the environment and backed by a major technology innovator, makes our Samasource workers feel valuable and important. It builds confidence and dignity.
And, importantly, the work provides key data to stop poaching. By tracking the location of elephants and screening for unauthorized vehicles and other signs of poaching, the Census provides meaningful data to the many organizations working on the ground in Africa. By themselves, these groups might be able to collect raw data but they’d not have the ability to glean much information from it, detect trends, or handle a large volume of images quickly.
The same techniques are being used to capture images from drones that monitor the oceans and other conservation areas. My brother Ved works with Stanford and NASA to image coral reefs from above and determine the degree of degradation. By mapping these corals over time, his team can determine the impact of global warming on coral bleaching, a major aspect of ocean health.
Diane Regas, who leads the Environmental Defense Fund, is at the forefront of ocean conservation and believes that drone monitoring is the future. Managing fish stocks requires stopping illegal fishing — but collecting this data can be costly and take a long time to reach relevant patrols. Real-time image tagging via drone or satellite completely changes the game. Imagine if local fishing patrols received alerts any time an illegal boat were discovered and could act on the spot.
Skycatch, founded by a friend and fellow kitesurfer Christian Sanz, collects this kind of data for construction and humanitarian relief efforts, tracking migrants across borders. Another fascinating company, Saildrone, co-led by my former board member Sebastien de Halleux, collects oceans data from autonomous sailing drones.
This is just the tip of the iceberg. Machine learning is changing every aspect of our relationship to the environment and allowing us to see patterns in real-time that would have taken years to understand in the past. It’s the same transformation that happened within healthcare with the quantified self movement.
At some point we won’t need humans to tag these images and generate meaningful data from them. Samasource will pivot into new types of work and move up the value chain. But for now, this work provides a lifeline to thousands of people in poor places, and it creates a more detailed picture of our impact on the planet.
Like this? Subscribe to my newsletter here to get regular tips and advice for social impact entrepreneurship.
