AI Governance Engineer
GitHub

Catastrophic forgetting

By Updated

The tendency of neural networks to lose earlier competence when trained on new tasks 1. It is a reason every retraining is a change event that re-runs the full eval suite, not only the tests for the new capability.

Developed in
ch. 11, Eight characteristics that break classic IT governance
Chapters
ch. 11, AI Defined
Source
1 numbered reference, listed below

Where it is used

One chapter of the Body of Knowledge uses the term. Each link opens the first section that does.

  • Dual use

    The capacity of the same AI capability to serve harmful ends as well as legitimate ones, for example a toxicity model inverted to propose toxic molecules.

  • Opacity

    The inability of a person to follow how a system reached an output.

Sources

  1. [1] Overcoming catastrophic forgetting in neural networks (networks lose earlier competence when trained on new tasks; arXiv 1612.00796). Kirkpatrick et al.. 2016-12-02. https://arxiv.org/abs/1612.00796 (verified: primary)

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

Cite this term

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

BibTeX

@misc{aige2026catastrophicforgetting,
  author  = {Jorge García Aibar},
  title   = {{Catastrophic forgetting}},
  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/catastrophic-forgetting}
}