Legal Research AI Evaluation Expert

About OpenTrain

OpenTrain is the #1 platform for finding and building careers in AI training and data labeling. We help people discover specialized AI training projects, build a lasting portfolio of work, and grow a durable freelance career in the fast-growing human side of AI.

• A focused hub for legal and technical AI training opportunities.
• Free to create a profile and apply for projects.

About AI training work

AI training (also called data labeling or human feedback work) is the practical, human side of building modern AI systems. Contributors review, correct, and rate model outputs so models learn to produce accurate, reliable results.

This work is a great fit for professionals seeking flexible, impactful contract work that directly shapes how next-generation AI behaves.

• Work uses real-world subject-matter expertise to improve model quality.
• Many projects are part-time and let you set weekly hours.

The role

OpenTrain is hiring a Legal Research AI Evaluation Expert to evaluate AI-generated legal research for accuracy, currency, and legal soundness. You will apply professional research standards to verify citations, analyze legal reasoning, and provide structured written feedback that helps improve model outputs on complex legal issues.

This is contractor, part-time work for US-based contributors with substantial legal research experience.

• Title: Legal Research AI Evaluation Expert
• Work type: Contractor, Part-time
• Location: United States only
• Time: 20+ hours per week

What you'll do

Your day-to-day work is hands-on review and evaluation of model-generated legal research. You will identify errors, assess whether authorities remain good law, and explain legal impacts clearly and concisely.

• Review AI-generated legal research against professional standards.
• Verify citations, holdings, statutes, and regulations for accuracy.
• Use KeyCite or similar tools to confirm good-law status and flag negative treatment.
• Identify missing controlling precedent and distinguish binding vs. persuasive authority.
• Provide structured written feedback and calibration notes to improve model reasoning.

Requirements

You must bring demonstrable, current legal research skills and the tools used by professionals to validate authority and citations.

• Substantive legal research experience in litigation, research attorney roles, or law librarian work.
• Regular use of Westlaw in current or recent professional work.
• Ability to validate case law, statutes, regulations, and good-law status with KeyCite or equivalent.
• J.D. and bar admission, or an MLIS (or equivalent) with documented legal research experience.
• US-based and able to work at least 20 hours per week as a contractor.

Preferred qualifications

The project favors candidates with advanced or practical research roles and multi-jurisdictional experience.

• Experience as a judicial law clerk, research attorney, or multi-jurisdictional research background.
• Comfort producing concise, defensible written explanations of research findings and authority treatment.

Who should apply

Apply if you regularly perform professional legal research, verify citations with Westlaw/KeyCite, and can explain legal rules and authority treatment clearly. This role is best for experts who enjoy improving model reasoning and seeing their feedback influence AI outputs.

• Experienced litigators, research attorneys, and law librarians.
• Applicants who can work independently and deliver high-quality written evaluations.

Pay and time commitment

This role pays hourly on a contractor basis. OpenTrain lists a rate range and expects contributors to commit to regular weekly hours.

• Pay: $75–$150 per hour (hourly pay; top rate $150/hr reflected in posting).
• Minimum commitment: 20+ hours per week.
• Employment type: Contractor, Part-time (US only).

How it works

You will evaluate text outputs from frontier models, assign ratings, and provide free-text corrections and rationale. Work is delivered through OpenTrain's project tools and guidelines; feedback is structured to help calibration and model improvement.

• Label types: evaluation ratings and text generation feedback.
• Data type: text — verify citations and legal reasoning in model outputs.
• You will receive project-specific instructions and examples to ensure consistency.

How to apply

If you meet the requirements and want to shape how AI handles legal research, apply with your OpenTrain profile and a brief summary of your legal research experience, Westlaw/KeyCite usage, and current bar or MLIS credentials. Qualified candidates will receive project onboarding and calibration tasks.

• Include your jurisdiction(s) and a short description of recent Westlaw/KeyCite use.
• Be prepared to complete calibration evaluations before work begins.