There’s work underway to democratize access to artificial intelligence-for-the-energy transition.
IBM, part of the computing old guard, is partnering with the United Nations Development Programme on two predictive AI models for energy and development, specifically in the Global South: the Electricity Access Forecasting model and the Clean Energy Equity Index.
The models, intended to give policymakers the tools they need to advance sustainable development, are a rare example of advanced models that are free and accessible to the public.
The Electricity Access Forecasting model forecasts electricity access through 2030 across 102 countries in Africa, Latin America, the Asia Pacific, and the Middle East. Using IBM’s software products such as the commercial generative AI and scientific data platform watsonx as well as an open-source machine learning library, it evaluates population, infrastructure, urbanization, elevation, and satellite land use data to make its predictions.
Meanwhile, the Clean Energy Equity Index combines geospatial analytics with environmental and social factors such as greenhouse gas emissions and relative wealth to highlight clean energy development needs and opportunities across 53 African countries.
The two models, which are available online, are intended to “provide a credible evidence base to help countries make meaningful and practical progress towards a just energy transition,” according to Laurel Patterson, who heads the UNDP’s sustainable development goals integration team. They’re part of UNDP’s Data Futures Exchange’s GeoHub, a collection of geospatial data and services.
The release is the latest in a recent slew of AI-powered tools that offer efficiency-through-information to the energy sector. Companies have been using AI models to support developers in choose locations for clean energy projects, collect heat data to implement cooling interventions and help utilities manage an alarming surge in power demand. However, the majority of those tools focus on North America or Europe — and access to valuable energy data tends to be paid-for.
When it comes to electrification and grid management, foundation AI models trained on vast amounts of data are considered particularly valuable, as they have the potential to improve grid planning, permitting, operations, and resilience.
