Sustainability & AI
Data labelling
Annotating data so models can learn from it, often performed by large low-paid workforces — a social dimension of AI supply chains.
Definition
Annotating raw data — tagging images, transcribing audio, categorising text — so that supervised machine-learning systems can learn from it.
Quick reference
At a glance
- Subject
- Sustainability & AI
- Editorial status
- Editorial draft
- Definition status
- Established
- Last updated
- 21 August 2026
- Also known as
- data annotation
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
Labelling is human work at industrial scale, often performed by distributed workforces on digital platforms under varying pay and conditions; the ILO has documented this segment of the AI supply chain. Label quality and consistency directly bound model quality.
How it is used
Environmental AI depends on it: species identifications for camera traps, land-cover tags for satellite tiles, relevance judgements for document classifiers. Some programmes use expert or community review; citizen-science platforms such as iNaturalist generate labelled data through participation.
Why it matters
Data labelling is where the 'artificial' in AI quietly becomes human. For a sustainability dictionary it matters twice: as a working-conditions issue inside the AI value chain, and as the epistemic foundation of every supervised environmental model.
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.