You Can Be Anything › All careers › Miscellaneous & Unique Professions
Ethical AI Trainer
Trains AI systems by providing feedback, labeling data, and testing for bias and safety.
Did you know?
AI trainers helped teach ChatGPT by having conversations with early versions and rating the responses
How subjects help
- Computer Science: Label data and test how an AI responds to find mistakes and unsafe answers.
- Psychology: Notice when an answer could mislead, upset or pressure the person reading it.
- Linguistics: Judge whether an answer is clear, natural and correct in grammar and meaning.
- Philosophy: Think through what fair and honest behaviour should look like for an AI helper.
A famous name
Joy Buolamwini: Tested AI face-recognition systems and showed in 2018 that they failed more often on darker-skinned women.
Try it yourself: Check a chatbot's answers
- Which answer is fair and correct?
- Which answer is unfair to some people, and which is simply wrong? Explain why.
- Rewrite the bad answers so they are correct and fair to everyone.
- Why is it important that someone checks what AI tools say before people rely on them?
Ethical AI trainers check chatbot answers for mistakes and unfair assumptions, and teach the system better answers.
Also try: Think through a big question.
Known for it in
USA, UK, Kenya, India, Philippines