By Michael Hainsworth
Artificial Intelligence is being deployed to protect the environment by cleaning up abandoned mines, determining the best locations for electric vehicle charging stations, and keeping cheating appliances off the electrical grid. At Natural Resources Canada, Vik Pant is leading the charge, but he can’t do it alone. “That’s not something that can be done humanly at any scale,” he says. There’s simply too much data to wade through. “We have access to satellite imagery. We have access to aerial photography from drones, from hot air balloons, from aircraft,” the chief scientist recounts. And much of that data comes from the public sector, space agencies, but also citizen scientists and university researchers.
“We can come together, work together and help each other solve these big problems.”
Vik Pant, Chief Scientist, Natural Resources Canada
Using machine learning to uncover patterns in the data, Dr. Pant is able to use these sources to uncover a mine abandoned decades earlier that may be leeching hazardous chemicals into the neighbouring water table. “You can train a machine learning model and then run that on your data set and get an up-to-date map of abandoned and active mines, which then you can update through time and over time, much faster than you’d be able to (manually).” Operating mines are also leveraging AI to reduce their impact on the environment using the internet of Things, devices and sensors connected via 5G and other wireless technologies. “Decisions are being made directly on that device or when it should cycle on when it should cycle off based on all kinds of things that a human would otherwise have to be doing. So that’s one part of it. The other part of it is the digital twinning.” By leveraging huge amounts of data from mining equipment and environmental sensors, AI researchers can digitally recreate a mine then run simulations of the impact certain drilling and remediation activities will have in the real world.
Meantime, in the real world, manually searching the internet for companies abusing the Energy Star logo is also something NRCan simply doesn’t have the human resources to do. Instead, NRCan is developing a spider that crawls through the internet using a variety of artificial intelligence technologies including image recognition to find companies falsely promoting their appliances as energy efficient. The Energy Star program is a North America-wide initiative that started in 1992. It indicates the expected energy consumption of a given appliance over a year and what that would likely cost. For the average homeowner, knowing this has limited value, but for a condominium developer who needs to buy hundreds of appliances, having the facts about how much it will cost to run a kitchen is only half of the solution. They need to show an increasingly environmentally conscious buyer that they understand the importance of minimizing a development’s carbon footprint. Appliances deemed environmentally friendly 30 years ago would no longer get that silver star of approval today.
“I think about sustainability at two levels. One is conscientious stewardship. The other is responsible consumption. When we talk about sustainability, you can think about it in many ways, but the way I like to think about it is effectiveness and efficiency. And we can think about effectiveness as doing the right things and efficiency is doing things right,” says Dr. Pant. Those same Millennial condo dwellers concerned about their carbon footprint are more inclined to purchase electric vehicles. The explosion in demand for EVs is tied largely to reducing “range anxiety” about running out of battery before arriving at Point-B, and nothing reduces the risk of an electrical panic attack than knowing there’s a charging station nearby. Dr. Pant’s team has been actively trying to determine the best locations for a Canada-wide network, and simply spacing them equidistant to each other isn’t efficient at all. EV charging demand is only one of the data sets that NRCan is using to determine the most sustainable locations. AI’s real power comes from combining seemingly unconnected data to come to more intelligent conclusions like weather patterns, commute times, holiday vacation schedules, and for Dr. Pant, the sources of electricity generation. “When you put data sets together, a whole new kind of data, potential gets unlocked, value potential gets unlocked,” he says.
It’s an interesting problem to build and assemble different types of machine learning models – some used to predict driver charging behaviour and others to ensure that the production, transmission, and distribution of electricity to those charging points is viable through time. Finding that intersection is what NRCan’s project helps to solve. NRCan sees leveraging a proliferation of sensors, probes, and instruments from the point at which electricity is generated to where it’s being consumed. If you’ve got a smart thermostat at home, you’re part of that big data pool being compiled by the maker of your internet of Things device. Bringing disparate data sets together often requires collaboration with the private sector. “We have world-leading scientists. We have world leading policy experts. But we’re not the only ones. There’s lots of universities, organizations like Microsoft and others that also have amazing data scientists. So, bring that knowledge pool together. Let’s align on mutual goals on our shared objectives and really create a win-win situation where we can come together, work together and help each other solve these big problems.”