The work
You will evaluate how an AI assistant uses information from a personal account to create relevant, accurate, helpful, and personalized responses. Your work will help improve the quality and reliability of AI systems.
You will follow project guidelines and apply the same evaluation standards across prompts and responses.
- Review AI-generated prompts and responses in English.
- Rate responses for relevance, accuracy, helpfulness, and personalization.
- Assess how well the assistant uses available account context.
- Write clear explanations and feedback for your ratings.
- Keep your evaluations consistent with the project guidelines.
What it pays and takes
This is a remote contractor project for applicants worldwide. Payment is in USD, but the listing does not provide a rate.
- Project length: 14 to 16 weeks, with possible extension opportunities.
- Time commitment: 20 or more hours per week.
- Language: Fluent English for reviewing prompts, responses, and account content.
- Account requirement: An actively used personal email account with at least one year of history.
- Photo requirement: An associated photo library with at least one year of content and activity.
- Usage requirement: Approximately 10 or more previous AI assistant queries or interactions.
- Settings requirement: English display settings for the relevant account services.
- Consent: You must be willing to sign a consent form.
- Helpful background: Experience reviewing AI outputs, evaluating prompts and responses, or providing structured written feedback.
How it works
Apply on OpenTrain with your resume, then complete the application on the hiring site.
About AI training work
AI training is the human work behind systems that generate and understand content. People review model responses, rate their quality, and provide feedback so AI can produce more useful results.