Lucas JoppaApr 22, 2019 18:49:07 IST
It doesn’t take computer vision to observe that humanity is already confronting a number of serious environmental challenges—from communities that are ravaged by floods and wildfires; to farmers who are losing their harvests to pests and extreme weather; to warming oceans, deforested wilderness, arid soil and contaminated watersheds.
As these crises unfold, they are devastating for local communities who go hungry, thirsty or lose their homes and livelihoods. But they also impact the broader ecosystem, rippling across national boundaries and disrupting the global economy. Extreme weather events (like hurricanes, droughts, and heatwaves) have all increased as a result of climate change and cost billions of dollars in business and property damage alone, not to mention loss of human life. Climate change is also transforming the global food supply by changing the migration patterns and availability of popular seafood species or causing unseasonable bumper crops in some regions, while entirely wiping out national cash crops in others.
These effects are particularly acute in poorer countries, which have lost billions of dollars a year in crops and livestock, and whose populations are most vulnerable to rising food costs. But food producers in places like France, Canada, India and the United States have also seen record-low yields and a negative influence on the quality, flavor, and nutritional value of their harvests.
In short, climate change is a massive problem across nearly every sector and measure of human development. To address it at the speed and scale that current conditions require, we’ll need to take a more data-driven approach — one that harnesses the full power of Artificial Intelligence (AI) and other advanced technologies to accelerate discovery and innovation at a truly planetary scale.
When world leaders agreed to adopt ambitious new targets for the reduction of global carbon emissions in 2015, it was a monumental diplomatic achievement and an encouraging step in the right direction for our planet. But the science shows that climate change is much more complex and policy levers that are focused on reducing CO2, alone, won’t save us from ecological collapse. We must do much more. We must radically transform how we monitor, model, and ultimately manage Earth’s natural systems, and quickly, in order to avoid “sleepwalking into a catastrophe.”
Though we face wave after wave of environmental crises, humanity is also fortunate to be riding a rising tide of technological innovation. Satellite imagery and remote sensing have brought more data to our fingertips than ever before, while powerful cloud computing and AI tools enable us to process that data into actionable intelligence and even forecast into the future. The challenge and, most importantly, the opportunity we have before us, is to deploy the advantages of machine learning and AI across every scientific field and industry, in order to chart a healthier and more sustainable path forward.
Accelerating research and innovation
In order to protect our planet, it is imperative that we first have a clear picture of how Earth’s systems work together and why its natural patterns are changing. For decades, and in some cases centuries, scientists have relied on the same tools and methods to observe the natural world — such as manually tagging and counting animals in the wild, or conducting pen and paper surveys of watersheds and forest lands. These methods are simply too slow and resource intensive to address the urgent needs of today.
AI and cloud technology can help catalyse scientific research, foster collaboration, and spur innovation on an unprecedented scale. As data pours in from a growing number of satellite images and remote field sensors deployed globally, researchers can now train machine learning models and use the massive computing power of the cloud to classify their data faster, and more accurately, than was previously possible. This saves organisations valuable time and money and reduces the time spent crunching data from years or months to days or even hours. It also frees up vital human resources to focus on developing new solutions. Successful models for tasks like species recognition or land cover classification can be shared as open-source APIs and adapted by researchers around the world to help solve similar problems, accelerating impact.
Driving data-informed decisions
Insights generated by the global proliferation of AI tools will also usher in a new era of more data-driven decision-making. Current policies — governing everything from carbon emissions and urban development to hunting and fishing regulations — are often based on datasets that are incomplete, imprecise and out of date.
In the United States, the best available land cover mapping data is just at 30-meter resolution and can be up to seven years old, because of the sheer amount of time it takes to manually label the high volume of new satellite imagery as it comes in. That means when local communities attempt to refresh their urban growth boundaries, or decide how much land to dedicate to agriculture, forestry, or habitat conservation, they’re often flying blind. The same is true on a global scale—in fact, in many countries good data is even more difficult to come by.
Cloud technology can help facilitate data and information sharing across national boundaries, which is critical because the environmental challenges plaguing our ecosystems don’t stop at the border. AI tools can help yield more accurate insights, empowering policymakers with more current and precise data to inform their planning. And well-trained machine learning models can be easily adapted across different datasets and geographies or used to run a variety of scenarios to test the impact of proposed interventions. This AI-enabled modeling is absolutely essential to developing policies and regulations that will help right the course for our planet.
Creating a green and flourishing economy
Another way that AI can enhance sustainability is in the economic arena. Many companies are already implementing new AI-enabled tools to help them cut costs, improve resource efficiency, and mitigate climate-related risks. For example, companies like Ecolab, Ørsted, and The Yield have developed AI-enabled solutions to improve water conservation, renewable energy management, and agricultural production. And a small startup called SilviaTerra is using AI to improve forest management and fuel the carbon offset market with precise data about the carbon sequestration potential of every acre of forestland in the United States.
These are not isolated examples. New research from PwC UK shows that AI can be applied to a wide range of economic sectors and industries to improve environmental outcomes and mitigate climate change — all while driving economic growth. In fact, PwC estimates that using AI for environmental applications could contribute up to $5.2 trillion to the global economy in 2030, a 4.4 percent increase relative to business as usual, and create an estimated 38.2 million new skilled, green jobs across the global economy.
This is the future of sustainability, and AI has a substantial role to play. Whether it’s deployed to monitor and improve environmental outcomes, enhance resource management or create a greener economy, taking a tech-first approach will upgrade our collective capability to be good stewards of the planet.
The author is the Chief Environmental Scientist, Microsoft and lead for the AI for Earth program
Tech2 is now on WhatsApp. For all the buzz on the latest tech and science, sign up for our WhatsApp services. Just go to Tech2.com/Whatsapp and hit the Subscribe button.