AI glossary · Basic · Safety & alignment · Defense & security
Dual-use technology
Dual-use technology is technology that can be used for peaceful civilian purposes as well as for military or harmful ones. Because most AI is general-purpose, this trait is especially strong in AI.
About this term
- Level
- Basic · Technical terms you often meet in AI news
- Category
- Industry AI
- Fields
- Safety & alignment
- Industries
- Defense & security
- Added
- 2026-10-10
How it works
Dual use describes the nature of a technology's uses, not how the technology itself works. For example, a model that finds objects in images can be used both for defect inspection in factories and for reconnaissance with satellite imagery. Large language models (LLMs) can help with summarizing documents and coding, but they can also be misused to build hacking tools or gather information on hazardous materials. So the same model can carry different levels of risk depending on who uses it and for what purpose.
Why it matters
Governments make dual-use technologies subject to export controls and investment screening. High-performance semiconductors and certain AI models and software often come up in these discussions. For companies, a product built for civilian use may face regulation depending on which country and which customer it goes to. In defense, dual-use technology is also a channel through which the military draws on fast-moving civilian AI progress, making it a core concept in defense procurement and civil-military technology cooperation.
Things to know
It is a balancing problem: restricting a technology also blocks its beneficial uses, while leaving it open raises the risk of misuse. For cases like open-weight models, which cannot be recalled once released, risk assessment and red-team testing before release become important. Developers need to decide how their terms of use handle military and surveillance uses, and customer verification and access control are also part of their responsibility. In defense use, whether human control remains over decisions to use force is treated as a core principle.
Dual-use checklist
Before supply
- Checked whether it is subject to export controls?
- Verified the customer and actual use?
- Done a risk assessment before release?
In operation
- Do the terms of use list prohibited uses?
- Is access controlled?
- Does human control remain over critical decisions?
Example
Suppose a company builds an AI that detects crop diseases from drone footage. With small changes, the same model could be used to find vehicles in military reconnaissance imagery. To sell the product abroad, the company needs procedures to check whether it is subject to export controls and to verify who the customer is and what the product will be used for.