AI Governance Engineer
GitHub

Regurgitation

By Updated

A model reproducing memorised training data verbatim, including personal data, whether prompted deliberately (training-data extraction) or not 1. Detected by output checks and canaries, and tested by extraction evals.

Developed in
ch. 19, AI-specific privacy breaches
Chapters
ch. 19, Privacy & AI · ch. 20, Existing Law
Contrast with
Hallucination
Source
1 numbered reference, listed below

Where it is used

2 chapters of the Body of Knowledge use the term. Each link opens the first section that does.

Regurgitation: mentions per chapter5 mentions of the term across 2 of the 23 chapters of the Body of Knowledge; the chapters the glossary cross-references are drawn solid.Cross-referenced chapter19 Privacy & AI: 3 mentions19 Privacy & AI320 Existing Law: 2 mentions20 Existing Law2No mentions in the other 21 chaptersSource: the Body of Knowledgev0.5.0
Regurgitation: mentions per chapter5 mentions of the term across 2 of the 23 chapters of the Body of Knowledge; the chapters the glossary cross-references are drawn solid.Cross-referenced chapter00010203040506070810111213141516171819 Privacy & AI: 3 mentions31920 Existing Law: 2 mentions220212223Source: the Body of Knowledge v0.5.0
Mentions per chapter How often each chapter names the term, with the chapters the glossary cross-references drawn solid. The longest bar is where to read the term in use; bars spread across many chapters mark a term that runs through the discipline.
Mentions as a table
Mentions of Regurgitation per chapter
ChapterMentionsCross-referenced
00 Preface0No
01 Definition0No
02 Why Now0No
03 Values & Principles0No
04 The Stack0No
05 Patterns0No
06 The Role0No
07 Maturity Model0No
08 Regulatory Map0No
10 Reading List0No
11 AI Defined0No
12 Governance Program0No
13 Risk Management0No
14 Development0No
15 Deployment0No
16 Fairness & XAI0No
17 Incidents0No
18 EU AI Act0No
19 Privacy & AI3Yes
20 Existing Law2Yes
21 AI Laws Worldwide0No
22 Principles & Standards0No
23 AI Agents0No
Regurgitation: related terms and patternsThe term at the centre; related terms on solid edges, the terms it is contrasted with on dashed edges, and the patterns that use it most in the colour of their layer.RelatedContrastRegurgitationRelated terms (5)Evals · RelatedEvalsMembership inference · RelatedMembership inferenceModel inversion · RelatedModel inversionPersonal data breach · RelatedPersonal data breachTraining, validation and testing data · RelatedTrainingContrast with (1)Hallucination · ContrastHallucination
Regurgitation: related terms and patternsThe term at the centre; related terms on solid edges, the terms it is contrasted with on dashed edges, and the patterns that use it most in the colour of their layer.RelatedContrastRelated terms (5)Evals · RelatedEvalsMembership inference · RelatedMembership inferenceModel inversion · RelatedModel inversionPersonal data breach · RelatedPersonal data breachTraining, validation and testing data · RelatedTrainingContrast with (1)Hallucination · ContrastHallucinationRegurgitation
The term's neighbourhood Terms related to this one on solid edges, terms it is often confused with on dashed edges, and the patterns that use it most in the colour of their layer. Follow a dashed edge to check the distinction before you use either term.
Neighbourhood as a table
Neighbourhood of Regurgitation
FamilyItemRelation
Related termsEvalsRelated
Related termsMembership inferenceRelated
Related termsModel inversionRelated
Related termsPersonal data breachRelated
Related termsTraining, validation and testing dataRelated
Contrast withHallucinationContrast
  • Membership inference

    An attack that determines whether a specific person's record was in a model's training set from the model's behaviour.

  • Model inversion

    An attack that reconstructs features of training subjects, such as a face, from a model's outputs and confidence scores.

  • Personal data breach

    A breach of security leading to the accidental or unlawful destruction, loss, alteration or unauthorised disclosure of, or access to, personal data, notified…

  • Training, validation and testing data

    The three data sets the AI Act defines for high-risk systems: training data fits the model, validation data tunes it and guards against overfitting, and…

  • Evals

    Automated tests of a model's or agent's behaviour (capability, safety and adversarial), run as controls, not as one-off research.

Sources

  1. [1] "Extracting Training Data from Large Language Models" (Carlini et al.; verbatim training sequences extracted, incl. personal data; arXiv 2012.07805). arXiv. 2020-12-14. https://arxiv.org/abs/2012.07805 (verified: primary)

Definitions of legal terms paraphrase the cited text, which governs. Dated statements are as of .

Cite this term

García Aibar, J. (2026). Regurgitation. In AI Governance Engineering: The Thesis & Body of Knowledge (v0.5.0), Glossary. https://doi.org/10.5281/zenodo.22956197. https://aigovernanceengineer.com/glossary/regurgitation. CC BY 4.0

BibTeX

@misc{aige2026regurgitation,
  author  = {Jorge García Aibar},
  title   = {{Regurgitation}},
  note    = {Glossary, AI Governance Engineering: The Thesis \& Body of Knowledge, version 0.5.0},
  year    = {2026},
  doi     = {10.5281/zenodo.22956197},
  url     = {https://aigovernanceengineer.com/glossary/regurgitation}
}