What is data annotation? Meaning, jobs and pay explained
By AI Global Work editorial team · Facts checked Nov 15 · live numbers update daily
Short answer
Data annotation means adding labels to data, such as images, audio or text, so that AI models can learn from it. As a job, it is remote contract work: you tag, transcribe, rate or correct data for a platform that pays by the hour or per task. Which platforms you can join depends on your country.
What data annotation means
An AI model learns from examples. Annotation is the work of preparing those examples: marking what is in a picture, writing down what was said in a recording, or saying whether an AI's answer was right. “Annotation” simply means adding a note or label; “data annotation” is doing that to data at scale. You will also see it called data labeling, and it is often misspelled “data anotation”.
What the work looks like
- Drawing boxes around objects in images or video.
- Transcribing audio, or checking a machine transcript.
- Sorting text into categories, or tagging names, places and intent.
- Rating or ranking AI answers against written guidelines.
- Recording your own photos, video or voice for a data-collection project.
Annotator, AI trainer or evaluator?
Platforms use these titles loosely and they overlap. Roughly: an annotator labels data; an AI trainer or tutor writes prompts and corrects a model's answers; an evaluator scores answers against a rubric. Specialist versions of all three ask for professional knowledge and pay more than general ones.
A job, a tool or a company?
- DataAnnotation (dataannotation.tech) is one company that pays contractors for this work. Its sign-in page is on its own site, app.dataannotation.tech. We are not DataAnnotation and can't log you in.
- Labelbox is a company that makes labeling software. It also runs Alignerr, its network of paid contractors.
- CVAT is an open-source annotation tool. It is software, not an employer.
- “Data annotation companies” or “services” usually means vendors that sell labeled data to AI developers. Contributor platforms are where individuals sign up to do the work.
What it pays
There is no single rate. Pay depends on the platform, the field and your country, and some roles pay per task. Our data annotation jobs page shows the live median of the rates platforms list. As published examples, DataAnnotation says its general projects start at $25 to $50 an hour, and Outlier lists US writing work at $15 to $35 an hour. Listed pay is not guaranteed income, because work depends on projects.
How to start
- Check which platforms accept your country before anything else. This differs a lot: DataAnnotation currently names six countries, while other platforms hire role by role.
- Apply on the platform's own website. Expect an assessment or a short AI interview.
- Never pay to apply, and don't buy an account. Real platforms don't charge.
- Apply to more than one platform, because projects come and go.
Questions people ask
What is data annotation?
Data annotation is labeling data, such as images, audio or text, so that AI models can learn from it. People are paid to do it remotely, by the hour or per task.
What does annotation mean?
Annotation means adding a note or label to something. In AI, it means labeling data so a model knows what each example shows.
Are data annotation jobs remote?
Almost always. The work is done online from your own computer. Each platform and role still decides which countries it accepts.
Is data annotation legit?
Yes, it is real paid work offered by established companies. Whether a particular offer is real is a separate question: apply only on a platform's own site and never pay a fee.
How much do data annotation jobs pay?
It varies by platform, field and country. DataAnnotation says general projects start at $25 to $50 an hour, and Outlier lists US writing work at $15 to $35. Our data annotation jobs page shows the live median of listed rates.
Can I do data annotation from my country?
It depends on the platform and the role. DataAnnotation currently names six countries, Outlier publishes a longer list, and other platforms decide role by role. Our country pages list the open roles that accept people where you live.
Do I need a degree for data annotation?
It depends on the platform. Outlier asks for at least an associate degree; DataAnnotation asks for a bachelor's degree or equivalent experience for general work. Specialist roles ask for more.
Where is the DataAnnotation login?
On DataAnnotation's own site, app.dataannotation.tech. AI Global Work is an independent site and has no login for it.