Chapter 09 · Sustainability & AISustainability & AI
Foundation model
Definition
A large-scale AI model trained on broad data that can be adapted, through fine-tuning or prompting, to a wide range of downstream tasks rather than built for one narrow purpose.
References
This source provides part of the technical or institutional basis for the definition.
This source supports the explanation of how the term is applied, measured or governed in practice.
Overview
What it means
The term was coined by Stanford's Center for Research on Foundation Models in 2021 to mark a shift from task-specific models to general-purpose bases such as large language and vision models. One model can now underpin thousands of applications, concentrating both capability and risk.
How it is used
In sustainability work, foundation models are adapted for report analysis, satellite-image interpretation and scientific text mining. Their training runs are also a major driver of AI energy demand, which is why the EU AI Act created specific obligations for general-purpose models.
Why it matters
A single base model shapes the accuracy, bias and footprint of everything built on it. For practitioners, foundation models lower the cost of applying AI to environmental problems; for policymakers, they concentrate accountability questions upstream.