Paid AI evaluation work in your language
AI models need people who know a language, a country or a field to check them. We're forming a small, paid, fully remote cohort, open to people in any country, who test AI systems in their own languages and areas of expertise: checking answers for accuracy, spotting harmful or biased output, and writing the hard questions models get wrong.
This is a waitlist. The cohort isn't running yet; we'll only start once the first project is funded, and we'll email everyone on the list before then.
What the work looks like
- Reading AI answers in your language and marking what is wrong, unclear or unsafe.
- Writing questions and scenarios from your country or profession that models handle badly.
- Checking translations and local facts in health, law, education and public services.
- Remote and part-time, a few hours a week, with written instructions and examples.
Our commitments
- Pay
- A published hourly pay floor before anyone starts, and payment for all accepted work.
- Adults only
- You must be 18 or older. We check this before the first project.
- No disturbing content
- The first projects won't include graphic or traumatic material.
- Your data
- We use your details only for this cohort and delete them if you ask.
For AI labs and research teams
If you need evaluation or red-teaming in specific languages, countries or professional fields, email [email protected]. We'll tell you who is on the list and what we can run.