skip to navigation skip to content

AI in energy policy


New paper co-authored by Dr Kamiar Mohaddes of Cambridge Judge Business School develops framework using artificial intelligence for better energy policy research in the Global South.

Smart city and a network.
Kamiar Mohaddes.
Kamiar Mohaddes

A new paper co-authored by Dr Kamiar Mohaddes, University Senior Lecturer in Economics & Policy at Cambridge Judge Business School, develops a new framework using artificial intelligence to improve energy policy research in the Global South.

Most energy policy research has been based on a clear separation between the natural sciences and the social sciences, and this leads to tech-focused bias and limitations in the definition of policy goals, says the paper published in the journal Energy Research & Social Science.

“‘Good’ energy policy under a technocratic directional bias thus becomes policy focused on ‘getting the technology right’. While technology can indeed be a force to achieve climate change mitigation and sustainability goals, its welfare effects may be restricted if the wants of a society are not appropriately reflected in policy”, the paper says.

The authors instead advance a “nested” framework based on artificial intelligence and machine learning-aided topic detection (known as “topic modelling”) that can reduce biases and lead to broader solutions by combining topic modelling, text-based narratives and grounded theory.

This approach “provides a higher degree of freedom to zoom-in, zoom-out and zoom-through the problem,” the paper says. “With zooming-out, possibilities and assumptions that are forgotten or are taken for granted can be better recognised; zooming-in aids better understanding of the granularity.

“Most importantly, the ability to zoom-through in the computational model can aid in critically analysing narratives as to material reality, norms and practices that determine the energy culture.”

The new framework analyses behaviour and data on energy use in order to better address energy and climate injustices in the Global South.

The research focuses on government-sponsored housing developed in Mumbai to house people living in slum dwellings as part of the Indian government’s poverty alleviation efforts. However, many people living in such slum rehabilitation housing (SRH) have been pushed into energy poverty due to rising electricity bills.

The framework analysis, by applying probability values, helped identify several factors that policy-makers might address, including irregular billing, lack of open spaces for children to play, and new home appliances bought with borrowed money that increases household debt levels. “Application of the framework revealed latent links between energy use and the built environment in the SRH that influence high energy bills,” the paper says.

The paper, which was supported in part by the Bill and Melinda Gates Foundation through the Gates Cambridge Scholarship, is entitled “Grounded reality meets machine learning: A deep-narrative analysis framework for energy policy research”.

The paper is co-authored by Ramit Debnath, Dr Ronita Bardhan and DrMinna Sunikka-Blank of the Department of Architecture at the University of Cambridge, Dr Sarah Darby of the Environmental Change Institute at the University of Oxford, and Dr Kamiar Mohaddes of Cambridge Judge Business School. Ramit Debnath and Dr Kamiar Mohaddes are both members of the Energy Policy Research Group at the Cambridge Judge Business School.