On October 20, 2023, Brian Plancher, assistant professor of computer science, published new co-authored research in the journal Communications of the ACM, titled “Is TinyML Sustainable? Assessing the environmental impacts of machine learning on microcontrollers.” The paper aims to address the environmental impacts of machine learning on low-power microcontrollers (TinyML).
In the article, Plancher and his colleagues assess the potential of TinyML to address the sustainable development goals outlined in the United Nations’ 2030 Agenda for Sustainable Development. The researchers assert that using machine learning (ML) on low-cost microcontrollers can unlock a whole host of sustainable applications ranging from wildlife monitoring for conservation to reductions in energy consumption through smart HVAC systems. However, the production and use of these devices also has an environmental cost. As such, the researchers developed an open-source Life Cycle Analysis (LCA) tool designed specifically for TinyML systems, and leveraged a scalability study to evaluate the net environmental impact of these devices.
Overall, Plancher and his colleagues are optimistic about TinyML’s sustainability potential. However, they emphasize that the realization of these positive results relies on the establishment and implementation of appropriate incentives and policies to ensure that sustainable applications are developed and sustainable practices are leveraged in the production and use of this technology.