Artificial Intelligence and Nuclear Command and Control
Abstract
An IISS tabletop exercise showed how an AI arms race could reduce strategic stability as nuclear-weapons states become more reliant on artificial intelligence.
Integrity signals
Registry
Cites
Statistics
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
Satellite images of the South China Sea posted online revealed the whereabouts of three Chinese submarines.
“In round one, reportedly from sources connected to QuantumAI, satellite images of the South China Sea were posted online, rendering the water transparent and revealing the whereabouts of three Chinese submarines.”
From Artificial Intelligence in the Military: An Overview of the Capabilities, Applications, and Challenges (2023) · cites this paper as [114]
Inputs can potentially lead an AI system towards an incorrect picture classification by slightly changing the pixels or patterns of the image.
“Tese inputs can potentially lead an artifcial intelligence system towards an incorrect picture classifcation by slightly changing the pixels or patterns of the image-changes that may be invisible to human sight and lead to unfortunate circumstances [114].”
From Artificial Intelligence in the Military: An Overview of the Capabilities, Applications, and Challenges (2023) · cites this paper as [114]
AI-driven management software at four commercial nuclear power reactors in heavily populous Chinese regions started reporting falsified safety alarms.
“Few years ago, the AI-driven management software at four commercial nuclear power reactors in heavily populous Chinese regions started reporting falsifed safety alarms.”
From Artificial Intelligence in the Military: An Overview of the Capabilities, Applications, and Challenges (2023) · cites this paper as [114]
Undermining the second-strike potential offered by stealthy nuclear-armed submarines represented a risk to strategic stability.
“Undermining the second-strike potential ofered by stealthy nuclear-armed submarines represented a risk to strategic stability [114].”
From Artificial Intelligence in the Military: An Overview of the Capabilities, Applications, and Challenges (2023) · cites this paper as [114]
A US Navy destroyer engaged in a freedom-of-navigation exercise in South China Sea territorial waters claimed by China came dangerously close to bumping with another PLA Navy vessel due to serious problems with its automated navigational systems.
“In the case of a fctional build-up with a scenario of crisis, a destroyer of the United States Navy engaged in an exercise called freedom-of-navigation in South China Sea territorial waters claimed by China came dangerously close to bumping with another vessel of the PLA Navy due to serious problems with its automated navigational systems.”
From Artificial Intelligence in the Military: An Overview of the Capabilities, Applications, and Challenges (2023) · cites this paper as [114]
DARPA awarded contracts to the US technology company QuantumAI for its cutting-edge magnetometers and gravimeters.
“Te Defense Advanced Research Projects Agency awarded contracts to the US technology company QuantumAI for its cutting-edge magnetometers and gravimeters.”
From Artificial Intelligence in the Military: An Overview of the Capabilities, Applications, and Challenges (2023) · cites this paper as [114]
GaiaForce could not locate any functional issues with the code of the AI-driven management software at the nuclear power reactors.
“GaiaForce could not locate any functional issues with the code.”
From Artificial Intelligence in the Military: An Overview of the Capabilities, Applications, and Challenges (2023) · cites this paper as [114]
The exercise showed the potential for another nonstate actor to steal private-sector technology advancements for malicious purposes, with disastrous repercussions for strategic stability.
“One of the most eye-opening faws the exercise showed was the potential for another nonstate actor to steal private-sector technology advancements for malicious purposes, with disastrous repercussions for strategic stability.”
From Artificial Intelligence in the Military: An Overview of the Capabilities, Applications, and Challenges (2023) · cites this paper as [114]
Cited by
Sorted by citations
Rashid, A. B., Kausik, A. K., Al Hassan Sunny, A., & Bappy, M. H. (2023) · International Journal of Intelligent Systems · doi:10.1155/2023/8676366 · Cited by 143 · see citing works
Horowitz, M. C., & Kahn, L. (2024) · International Studies Quarterly · doi:10.1093/isq/sqae020 · Cited by 68 · see citing works
Artificial Intelligence in Nuclear Warfare: A Perfect Storm of Instability?
Johnson, J. (2020) · The Washington Quarterly · doi:10.1080/0163660x.2020.1770968 · Cited by 50 · see citing works
Reciprocal trust and distrust in artificial intelligence systems: the hard problem of regulation
Maggetti, M. (2026) · AI & SOCIETY · doi:10.1007/s00146-026-02966-2 · Cited by 2 · see citing works
- Authors
- Fitzpatrick, M.
- Corresponding author
- Fitzpatrick, Mark
- Venue
- Survival
- Publisher
- Informa UK Limited
- Published
- 2019-05-04
- Volume
- 61
- Issue
- 3
- Pages
- 81-92
- ISSN
- 0039-6338 (ISSN-L)
- 1468-2699
- OpenAlex
- W2946716844
- Cited by
- 26
- References
- 0
- Citations by year
- 2026: 1 · 2025: 7 · 2024: 4 · 2023: 5 · 2022: 2 · 2021: 3 · 2020: 3 · 2019: 1
- FWCI
- 9.52
- Type
- journal article
- Language
- en
- Open access
- closed
- Locations
- Survival (publishedVersion)
- Funders
- Carnegie Corporation of New York
- Defense Advanced Research Projects Agency
- Topics
- Social Sciences › Political Science and International Relations › Nuclear Issues and Defense
- Physical Sciences › Engineering › Control and Systems Engineering › Military Strategy and Technology
- Social Sciences › Political Science and International Relations › European and Russian Geopolitical Military Strategies
- Keywords
- artificial intelligence, nuclear command and control, strategic stability, nuclear arms race, nuclear deterrence, tabletop exercises, nuclear weapons
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review