AI Governance Engineer
GitHub

Dual use

By Updated

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 1. It is answered with misuse threat models, red-team cases for harmful uses of legitimate capability and runtime detection.

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

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

  • Catastrophic forgetting

    The tendency of neural networks to lose earlier competence when trained on new tasks.

  • Opacity

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

  • Threat model (AI)

    A versioned record of what can go wrong with an AI system and what is done about it: data flows and trust boundaries, threats per element from STRIDE and…

Sources

  1. [1] "Dual use of artificial-intelligence-powered drug discovery" (an inverted toxicity model proposed about 40,000 candidate toxic molecules in under six hours; Nature Machine Intelligence). Urbina, Lentzos, Invernizzi and Ekins (PubMed Central). 2022-03-07. https://pmc.ncbi.nlm.nih.gov/articles/PMC9544280/ (verified: primary)

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

Cite this term

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

BibTeX

@misc{aige2026dualuse,
  author  = {Jorge García Aibar},
  title   = {{Dual use}},
  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/dual-use}
}