Fine Tuning
Adjusting an AI model by training it further on a set of examples, so that it writes, formats or classifies in a consistent way afterwards. It solves a behaviour problem rather than a knowledge problem, and what it learns is absorbed into the model itself.
What it means in practice
The RAG and fine-tuning comparison sets out the usual order: clearer instructions first, then retrieval-augmented generation, and fine-tuning only for a narrow, repetitive task that retrieval cannot reach. A fine-tuned model goes stale until it is retrained and is hard to audit, while a document store can be updated or cleared directly. The two can be combined, with fine-tuning for tone, format or a narrow classification and RAG for the facts.
How we use it
The same comparison flags the data protection difference: data used to fine-tune a model is absorbed into it, and the EU data protection board has said whether a trained model counts as anonymous must be judged case by case. A document can be removed from a RAG store in one checkable step; a fine-tuned model has no equivalent, which is why what goes into training needs deciding before it is used, a data protection question as much as a technical one.
FAQ
Questions about fine tuning
What is the difference between Fine-Tuning and RAG?
Fine-tuning changes the model by training it on examples, so it behaves consistently. RAG leaves the model unchanged and looks up your documents when a question is asked. The RAG and fine-tuning comparison puts it as a behaviour problem against a knowledge problem.
Does Fine-Tuning stop an AI from making things up?
Not reliably. The comparison cites a 2023 study in which RAG outperformed fine-tuning for both existing and new knowledge, and a fine-tuned model can still give a confident wrong answer.
What it means for a Business
Most business needs are knowledge problems, such as staff asking about policies, contracts or products, and those are solved by giving the model the documents rather than retraining it. Fine-tuning earns its place when one narrow, repetitive task needs identical behaviour thousands of times, and only after clearer instructions have been tried.
Ready to put this to work?
Tell us where your team is with AI and we will tell you honestly what would make the biggest difference.