---
title: "Regurgitation"
description: "A model reproducing memorised training data verbatim, including personal data, whether prompted deliberately (training-data extraction) or not."
canonical: https://aigovernanceengineer.com/glossary/regurgitation
author: "Jorge García Aibar"
license: "CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/)"
doi: https://doi.org/10.5281/zenodo.22956197
version: "0.5.0"
updated: 2026-09-24
---

# Regurgitation

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](https://aigovernanceengineer.com/bok/privacy-and-ai#ai-specific-privacy-breaches)
- Chapters: [ch. 19, Privacy & AI](https://aigovernanceengineer.com/bok/privacy-and-ai) · [ch. 20, Existing Law](https://aigovernanceengineer.com/bok/existing-law)
- Contrast with: [Hallucination](https://aigovernanceengineer.com/glossary/hallucination)
- In the glossary chapter: https://aigovernanceengineer.com/bok/glossary#t-regurgitation

## Sources

[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)