Foundation model
A large general-purpose model that can be adapted to many tasks.
Foundation models are trained once, at great cost, often by a third party, then adapted through prompting, fine-tuning or retrieval. The organisation using one inherits its strengths, limitations and biases.
On the exam
Using a third-party foundation model is an AI supply chain risk: assess the provider’s documentation, testing, data practices and contract terms.
Related terms
- Fine-tuningFurther training a pre-trained model on specific data to adapt it.
- AI supply chain and third-party AIRisks from models, data and AI services obtained from outside.
- Model cards and datasheets for datasetsStandard documentation describing a model or a dataset.
AAIA Test yourself on this with timed AAIA practice questions in the AAIA Exam Simulation Hub, or read the free AAIA exam guide.
Test yourself with today’s question of the day.
