The process of adapting a pre-trained AI model to perform specific tasks by training it on additional, domain-specific data while preserving its foundational capabilities.
Definition
Fine-tuning is the process of further training a pre-trained artificial intelligence model on additional, domain-specific data to specialize its capabilities for particular tasks, domains, or use cases. This technique allows organizations to leverage the general knowledge encoded in large foundation models whilst customizing them for specific applications, creating a more targeted and effective system for particular purposes.
Fine-tuning represents a critical juncture in AI system development where general-purpose models are transformed into specialized applications, often crossing from general-purpose platforms into specific use cases subject to sector-specific regulations and liability frameworks. The process typically involves taking a pre-trained model and continuing its training on a smaller, curated dataset relevant to the intended application.
In legal contexts, fine-tuning has gained significance as it represents a key point where responsibility and liability may shift from foundation model providers to downstream developers and deployers. Understanding fine-tuning is essential for addressing questions of regulatory compliance, professional responsibility, and liability allocation in AI applications.
Regulatory and Compliance Implications
Fine-tuning often transforms general-purpose AI models into systems subject to specific regulatory requirements based on their intended use. A foundation model fine-tuned for medical diagnosis becomes subject to medical device regulations, whilst one adapted for employment screening must comply with anti-discrimination laws and fair hiring practices.
The EU AI Act's risk-based approach means that fine-tuning may elevate a model's regulatory classification. A general-purpose model with minimal regulatory obligations may become a high-risk AI system after fine-tuning for specific applications such as credit scoring, law enforcement, or healthcare diagnosis, triggering comprehensive compliance requirements.
Sources
European Union, Regulation (EU) 2024/1689 laying down harmonised rules on artificial intelligence (Artificial Intelligence Act), Official Journal of the European Union, L 1689, 12 July 2024. Various industry and regulatory sources addressing fine-tuning practices and legal implications.
Example
Fine-tuning a general-purpose model on legal contracts could produce a specialised document-review assistant.