Sustainability & AI
Black-box model
An AI model whose internal reasoning is not interpretable by its users, complicating accountability for its outputs.
Definition
A model whose internal reasoning is not interpretable to its users or developers, so that outputs can be observed but not straightforwardly explained.
Quick reference
At a glance
- Subject
- Sustainability & AI
- Editorial status
- Editorial draft
- Definition status
- Established
- Last updated
- 21 August 2026
References
This source provides part of the technical or institutional basis for the definition.
This source supports the explanation of how the term is applied, measured or governed in practice.
Overview
What it means
Deep networks with millions or billions of parameters are the archetype: their decision logic is distributed and opaque. 'Black box' is a property of the model-human relationship, not of accuracy — an opaque model can be highly accurate and still resist explanation.
How it is used
Opacity complicates every sustainability use where reasons matter: an AI-screened supplier, a scored credit, a flagged claim. Explainability techniques and documentation practices exist to open the box partially, and the EU's trustworthy-AI framework lists explicability as a core requirement.
Why it matters
When an unexplainable model informs environmental or social decisions, accountability has nowhere concrete to land. Black-box risk is therefore a governance category, not just a technical footnote — it determines where AI can responsibly be used at all.
Review
Help keep this definition useful and accurate.
Submitted reviews stay private until an editor decides whether to accept and attribute them.
Endorse this definition
Confirm what works
Endorse the current wording when it is accurate and useful. Suggested changes use the separate editorial form below.