From healthcare to manufacturing, clean technologies, finance and many other industries, there is growing hype around artificial intelligence (AI) and its potential to revolutionize operations.
The increasing allure of AI was evident in this year’s federal budget, where the government announced that it would invest $2.4 billion to advance AI development and adoption in Canada.
While AI offers the promise of transformational leaps in efficiency, productivity and innovation, it is not without significant costs, including to the environment. The world is in the midst of a climate emergency, making it imperative that we temper our excitement around AI with a sober reflection on its impact on climate change.
Greenhouse gas emissions continue to warm the planet, breaking rising global temperature records and increasing sea levels at unprecedented rates as ice cover declines and glaciers retreat. Extreme weather events, floods, droughts and wildfires have become more common, with devastating consequences, including in Canada.
At the same time, there is a technological revolution underway as AI becomes more pervasive. Its ability to quickly analyze vast amounts of data, recognize patterns and automate repetitive tasks is driving government, business and industry interest in it.
And as countries — including Canada — strive to meet their climate goals, some see AI as an indispensible tool. Its algorithms can build more accurate climate models, provide better ice mapping, improve energy efficiency in buildings and decarbonize industrial sectors, among other uses.
Yet, AI is not the cure-all. It consumes vast quantities of energy as it develops and runs the algorithms it uses to analyze data and recognize patterns. If the energy comes from fossil fuels, it directly contributes to carbon emissions.
The computer equipment needed for AI must be regularly cooled to keep the machines operating, requiring large volumes of water while the world faces a freshwater scarcity. Discarded computer hardware — including waste from AI data centres — only adds to the planet’s e-waste problems.
We also cannot overlook the fact that many AI applications focus on mitigation and adaptation, not the root causes of climate change. This is an important shortcoming. Indeed, scientists have expressed frustration that social and political inaction — not technical knowledge or means — impedes us from tackling the root causes, the burning of fossil fuels and depletion of forests.
In engineering, we teach students to look at problems holistically. We have many engineering case studies that show the importance of thinking about how to set up a problem and not just rushing into solving what we think the problem is. We need to use this thought process in relation to AI and climate change.
For instance, is the climate emergency a technical problem or a political one and can AI help if it is a political problem versus the perspective that AI will fight climate change. How we set up the problem is critical to how we solve it.
Enthusiasm for AI must not lead us to skip the important work of “problem setting.” Jumping right to solving a problem without carefully defining it could overlook key issues. When it comes to AI and the climate emergency, we should consider a full range of questions that will shape what problems we think we ought to be solving.
For instance, we need to consider what we already know about climate change, about how to slow climate impacts and what communities need. Does a community want to know how to predict bigger storms through AI intervention or does it need resources for deep retrofits or making buildings more resilient?
We also need to recognize that we cannot rely on AI alone to save us from climate change. We need to tackle the problem holistically and utilize trained experts. While AI specialists may understand the technology, they are not necessarily experts in climate change.
We must continue to make room for the trained experts in climate science who can make decisions using all of the facts and difficult contextual data, including whether a particular solution is in line with the values of a community.
Instead of jumping head first into AI, policymakers and decision leaders need to consider all of its ramifications. After all, the decisions we make today about where to allocate resources — technical, environmental and social — will shape the decisions we can make in the future.
Photo courtesy of DepositPhotos
