For Heads-of · Practitioner
Training Data
The dataset used to train a machine learning model, from which the model learns patterns and relationships.
- training data
- data quality
- model training
- dataset
Quality impact
A model's performance and behaviour depend heavily on its training data. Biased, incomplete or low-quality data produces a biased or unreliable model, the register's training data bias and discriminatory outcomes.
Data governance
Training data should be documented: source, collection method, size, representativeness, known biases, licence and lawful basis, and how it was cleaned or filtered. Training data provenance checks keep that record.
Regulatory scrutiny
The EU AI Act requires data governance for high-risk systems (Article 10) and a public summary of training content from general-purpose model providers. Using personal, copyrighted or confidential data in training raises separate questions under data protection and copyright law.
Related Terms
Data Poisoning
An attack in which an adversary deliberately corrupts data an AI model learns from, to degrade it or plant hidden behaviour. It now reaches beyond training data to fine-tuning sets, retrieval corpora and agent memory, and a poisoned model can pass ordinary evaluation.
Data Quality
The degree to which data is accurate, complete, consistent and representative of what it is meant to measure.
Fine-Tuning
The process of adapting a pre-trained model to a specific task or domain by training it further on a smaller, task-specific dataset.
Synthetic Data
Artificially generated data, created by algorithms or models, used instead of or alongside real data in training and testing.