The Work
You will evaluate AI-generated responses to U.S. legal hypotheticals across several areas of law. Your reviews will help measure how well AI systems spot issues, apply legal rules, reason through problems, and reach accurate conclusions.
You will use a structured rubric, identify difficult edge cases and gaps in legal reasoning, and write detailed annotations and feedback. You will also work with research and project teams to keep evaluation standards aligned with project goals.
- Review AI-generated responses to U.S. legal hypotheticals.
- Assess issue spotting, legal accuracy, reasoning quality, and application of law.
- Identify gaps, blind spots, and edge cases in legal reasoning.
- Write detailed annotations and feedback for model improvement.
- Collaborate with research and project teams on evaluation standards.
What It Pays And Takes
This is a remote contractor project. The project requires 40 total hours over four weeks, with a stated time requirement of 20 or more hours per week.
- Pay: Up to $150 per hour, depending on internal evaluation and location.
- Project length: 40 total hours over four weeks.
- Work arrangement: Remote contractor project.
- Availability: 20 or more hours per week.
- Education: Juris Doctor degree from an accredited U.S. law school.
- License: Active admission to a U.S. bar.
- Experience: At least three years of U.S. legal practice.
- Skills: Strong reasoning across core areas of U.S. law, excellent written communication, careful attention to detail, and high ethical standards.
- Language: Professional fluency in English.
- Helpful background: AI annotation, model evaluation, structured feedback on generated responses, and analysis of complex legal hypotheticals.
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 learn from examples. Legal experts help by reviewing model responses, applying professional standards, and explaining where an answer is accurate, incomplete, or poorly reasoned; specialized expertise is important because it supports more reliable AI evaluation.