Set by the law · AI Act Art. 3(29)
Machine learning and AI training data
In plain English
Machine learning lets AI systems learn from training data. High-risk systems need quality training data under the AI Act, and the EDPB says models trained on personal data are not automatically anonymous.
The official text
‘training data’ means data used for training an AI system through fitting its learnable parameters
High-risk AI systems which make use of techniques involving the training of AI models with data shall be developed on the basis of training, validation and testing data sets that meet the quality criteria
the EDPB considers that AI models trained on personal data cannot, in all cases, be considered anonymous. Instead, the determination of whether an AI model is anonymous should be assessed, based on specific criteria, on a case-by-case basis.
Related
The official text is quoted word for word from the English version of the source named under it; only the version published in the Official Journal of the EU is legally binding. The plain-English line is Fino’s own summary, not legal advice.