Glossary · 1 · Foundations: how AI works
Foundation model
A foundation model is a large AI model trained on broad data at scale that can be adapted to a wide range of tasks — through prompting, retrieval or fine-tuning — instead of being built for a single purpose.
- Intermediate
- Technical project managers
- Developers
In one sentence
Foundation models defined: large, general AI models adapted to many tasks, and how they relate to the AI Act’s general-purpose AI models.
Example
One foundation model powers a company’s documentation chatbot, its translation workflow and its code assistant, each adapted with different prompts and data.
Why it matters on your learning path
- Technical project managers: Building on a foundation model shifts effort from model development to integration, evaluation and vendor management.
- Developers: Choose a foundation model by task fit, context size, cost, licensing and data terms — see model selection.
Foundation model vs. general-purpose AI model
“Foundation model” is a research term coined at Stanford in 2021. The EU AI Act uses the legal term general-purpose AI model, with specific obligations for providers.