AIRiskAware
AI Governance Glossary
Governance Practice

What Is Fine-Tuning?

Fine-Tuning is the process of further training a pre-trained AI model on a narrower, task-specific dataset to adapt it to a particular use case.

Definition

Fine-Tuningthe process of further training a pre-trained AI model on a narrower, task-specific dataset to adapt it to a particular use case.

Fine-tuning takes a general-purpose foundation model and specialises it — for example, adapting a base language model to legal document review or clinical note summarisation. From a governance perspective, fine-tuning is significant because it can introduce new risks (bias from the fine-tuning data, capability changes, degraded safety guardrails) and because it may shift regulatory responsibility: under the EU AI Act, an organisation that substantially modifies a GPAI model can become a provider with its own obligations.

Source: EU AI Act, Article 25; NIST AI 100-1

Plain-language explanation

Fine-tuning takes a general-purpose foundation model and specialises it — for example, adapting a base language model to legal document review or clinical note summarisation. From a governance perspective, fine-tuning is significant because it can introduce new risks (bias from the fine-tuning data, capability changes, degraded safety guardrails) and because it may shift regulatory responsibility: under the EU AI Act, an organisation that substantially modifies a GPAI model can become a provider with its own obligations.

Primary source: EU AI Act, Article 25; NIST AI 100-1

See where you stand on AI governance

Take the free 7-question maturity assessment and get a personalised action plan.

Free assessment — 3 minutes →