This session was recorded in NYC on October 22nd, 2019 and can be viewed here: https://youtu.be/_YUo5dGZqEM
Conservation & AI
Bio: Niraj Swami is an avid technologist & innovator with a passion for building solutions at the intersection of Artificial Intelligence, human-centered design and behavioral economics. Niraj has led AI and innovation strategy for enterprises, non-profits and startups in the human capital, healthcare and conservation industries. He enjoys exploring the AI-Human relationship with ventures that weigh on purpose, fairness and balance. Niraj is the founder of SCAD AI, a boutique purpose-tech venture firm, and a Senior Advisor for Applied AI & Innovation Ventures at The Nature Conservancy, a global non-profit solving the planet’s greatest conservation problems. Niraj holds an Honors MBA from the University of Chicago Booth School of Business and a Summa Cum Laude Software Engineering degree from Marquette University. When not tinkering with ideas, he enjoys writing music and traveling in search of orcas in the wild.
6. What if we weaved in science, insights & intelligence
into the daily fabric of scientists, conservationists,
decision-makers and even citizens?
Emerging technologies and tools like Artificial
Intelligence, Human-centered Design & Advanced Data
Science techniques can accelerate how our work is
measured & democratized.
What if we tech-enable our “how”?
Photo by Casey Horner on Unsplash
THE NATURE CONSERVANCY 6
10. • Going beyond satellite imagery: Capturing
new dimensions of ground truth with audio
sensors!
• Monitoring & evaluation using drone footage &
artificial intelligence!
• And many more in our pipeline...
More Initiatives
Photo by Erik Odiin on Unsplash
THE NATURE CONSERVANCY 10
11. THE NATURE CONSERVANCY 11
1 IMPACT
ENGINE2DATA
EXCHANGE
2 TECH-ENABLED FOCUS AREAS to help us harness tech for conservation.
12. THE NATURE CONSERVANCY 12
By scientists, for everyone
TOOL 1: DATA EXCHANGE
• Thousands of datasets
• Open access
• Expert-oriented, but not impenetrable
13. THE NATURE CONSERVANCY 13
Turning data & context into actionable insights
TOOL 2: IMPACT ENGINE
14.
15. With the Techstars Partnership we’ve already begun to
address technologygaps and practical applications.
One of the reasons I came to TNC is my deep respect for the organization’s science
TNC solves problems really well—for both nature and people
We’re good at determining where biodiversity needs to be protected and finding ways to conserve it
Whether that’s a marine protected area or a big national park
And we’re good at applying solutions that help people, like water funds to provide cities with cleaner water or coastal conservation that makes communities more resilient to rising seas
I think the best way to explain these tools is through example
Here’s a global problem that TNC is committed to helping solve: recurring water shortages
Half of our cities and three quarters of our farms experience shortages
This affects agriculture, it affects urban water supplies, and of course it affects nature
According to WWF’s Living Planet Index, freshwater species declined by a staggering 75% over the last 40 years
We are already using technology like artificial intelligence and machine learning to make our solutions even more effective
Essentially, AI means we rely on computers to use data to replicate human thinking—but reaching conclusions much more quickly
Right now, for example, we’re working with a bunch of partners to thin and use controlled burning to reduce the risk of catastrophic wildfires in California’s Sierra Nevadas
We’re testing an AI program designed to quickly assess whether our plan for thinning will leave enough open space to help prevent massive fires
So for this project, the AI program will use daily, high-resolution satellite imagery of pre- and post-thinning work in near real time—the AI not only collects the images, it reviews them and makes decisions about where to focus efforts
This saves fire crews from having to rely on expensive airplane overflights or less-accurate assessments from the ground
So AI is helping us do this critical work both faster AND more accurately
Another example is how we are using machine learning to measure the impacts of conservation projects
We already have a great low-tech solution. Decades ago, TNC created state-based natural heritage programs to measure the condition of animals and plant communities.
This data is used to justify tax credits for conservation easements, for example.
Every state has such a program, and to ensure consistency every one uses the same A-D grades for ecological condition.
The problem is that the data is collected by people going out into the field and counting things. It’s way too costly to do thing this way.
So we are now partnering with a start-up, to train a deep learning model (AKA MACHINE LEARNING) that will predict these data using satellite images.
We just got initial results last week for the first step using data from one plant community of high biodiversity importance, longleaf pine forests.
Our model can classify these pine forests with 78% accuracy. This is a big deal in and of itself because currently we cannot distinguish these forests from commercial pine forests that are not important for biodiversity.
The next step is to get additional data and train the model further to classify the ecological condition of different plant communities.
Imagine this info in the hands of conservation organizations and decision makers.
For the first time ever, we will be able to very quickly and very accurately evaluate project outcomes and inform future actions!
My job at TNC is to get us to more of this, and move at the speed of tech
I envision three BIG tools that will help us do that:
A data exchange, an analytics-driven impact engine and a product shop
So what do we do?
And when I say “we,” I don’t just mean TNC
I mean the world
Think of that Waze example
Waze offers you a better route to wherever you are going by analyzing data and making predictions
We want to put the planet on a better route
The first step is a data exchange
Thousands upon thousands of datasets, all open access and available to experts, all kept fresh and up-to-date
Publish data, publish data, publish data!
To solve water scarcity, we will capture remotely-sensed data from watersheds all over the world, and collect it in the exchange
Again, we is TNC, NGO partners, businesses, universities—the exchange will be place to hold any data that can contribute to a water solution
The second tool is an impact engine that translates the data into information - By using machine learning to do the analytics
The impact engine removes the expert requirement and lets lots of people access it fast
Again, we are changing the route – not for your trip to Starbucks, but for the planet’s water use
And telling TNC, for example, exactly where to put one of our own best tools: a Water Fund that is all about source water protection
What a water fund does is pays for upstream land conservation to improve water supply for downstream communities
But this is way bigger than TNC – with the right data and the right machine intelligence, the engine will be able to predict droughts better, so farmers can be proactive instead of reactive
It will give the international finance community better information to make investment decisions through simulations and long term forecasting
So how do we do that?
We started with Techstars—a partnership we’ve recently launched to leverage that entrepreneurial spirit that has disrupted so many industries
And put it to work for conservation
Techstars is really about partnering with start-ups and learning from them as well
We are not a technology company, obviously, but we absolutely see the value of this kind of partnership and we know technology is key to solving the world’s biggest problems
So that’s what we feel like we need to do to bring conservation into the present
And give the world’s problem solvers the data and tools they need to protect nature, tackle climate change, provide food and water sustainably and build healthy cities
That’s our vision: people using technology to solve conservation problems faster and more effectively