With a background in physics and mathematics, the assistant professor of public policy researches innovative technologies through the lens of climate change and related policy.
Anthony Harding worries that climate messaging is not always easily digestible for the lay audience. When scientists predict that mean global temperatures are going to rise by 1 ℃, for example, what will the impacts of such warming look like?
“Understanding what that actually means for people and what we can do about it is more in the economic and policy domain and that’s what interests me,” says Harding, Assistant Professor at the Carter School of Public Policy at Georgia Tech. Economic theory explains how economies grow and change over time, so factoring in climate change impacts can help create a model that predicts the effects of rising temperatures. It also enables economists to contrast the changes to the alternative: no global warming. From a policy perspective, we want to try to balance the costs of climate change policies against expected costs of climate change.
After an undergraduate degree in math and physics at Rensselaer Polytechnic Institute, Harding was keen on an interdisciplinary approach to technology. Economics appealed because it enabled Harding to pursue applied research. Harding received his doctoral degree in economics from Georgia Tech and pursued post-doctoral work at the Kennedy School of Government at Harvard where some of his work focused on the economic modeling of climate impacts. Having a physics background helped understand the science behind the climate models better.
At Georgia Tech, Harding uses applied economic theory and econometrics to understand climate impacts and related global policies.
Researching the energy impacts of AI
One of Harding’s research areas evaluates the energy impacts of AI technologies.
“That connects with my broader research themes of climate and economic theory and empirical methods in that energy use is one of the main drivers for climate change,” he says.
A paper based on this research explores the direct impact of AI on energy use such as that needed for processing models. In addition, it evaluates AI’s indirect effects on energy consumption.
“If AI makes me more efficient, I’m going to produce more things. If I produce more things, that is also going to require more energy, not just to power the AI, but also up and down the production network. So we need to adopt a more holistic picture to understand the overall aggregate impact of AI on energy use,” Harding says. One of the research goals is to evaluate how energy use in the economy might adapt given productivity changes due to AI.
Harding and team are calibrating an economic theory model using preliminary data to evaluate AI-related shifts in energy use across different sectors of the economy.
A surprising finding
Contrary to what we have been hearing on the news, Harding’s preliminary research finds the expected energy use related to AI over the next ten years to only be a fraction of a percent. Part of the explanation might lie in the possibility that reports might be conflating all data center use with AI loads, when that’s not always the case, Harding argues. “AI might have a large growth but it’s not the dominant share of current energy use in data centers,” Harding says. AI-related energy use is also a matter of scale and perspective, Harding adds. “Across the entire United States, the expected change in energy use from AI models might be pretty small but if you’re an individual city or state hosting these data centers, it might be a big deal for you,” he says.
Future research
Harding is an avid lover of houseplants and loves taking his dog on walks. As for what’s next in research, Harding is keen on exploring its effects on the local scale, in Georgia. Working with researchers to gather data on the location of data centers across the state and the rest of the United States, he is hoping to explore how local communities are perceiving the impacts of these data centers.
Broadly, Harding’s goal continues to be to use economic theory as a tool to explore innovative climate technologies and policy. Among these innovative technologies: solar geoengineering, which can include painting roofs white to injecting aerosols in the stratosphere.
Harding paints the promise of the technology with a lot of caveats. “If done in a very particular way and if well-deployed, solar geoengineering might be able to moderate a lot of expected harms due to climate change. For some of the most costly impacts of climate change, such as impacts on agriculture, human health, and economic growth, research suggests that people who live in poorer, hotter areas of the world will likely benefit the most,” he says.
News Contact
Story Written by: Poornima Apte
Priya Devarajan || Research Communications Program Manager