Legal expertise extends beyond knowing the law. Attorneys evaluate incomplete facts, reconcile conflicting authorities, and advise clients in ambiguous situations. AI's value will depend on how well it supports this reasoning.
Mercor matches law professionals with frontier labs building the next generation of legal AI systems. Earn $55–$150+/hour to help.
Your expertise helps shape how future models reason through complex legal questions involving ambiguity, risk, and high-stakes decision-making.
Compensation typically ranges from $55–$150+/hour depending on specialization, experience, and project complexity, with opportunities for ongoing project work.
Contribute from anywhere and choose projects that fit your schedule. Work independently without long-term commitments or fixed hours.
Go beyond legal research and document review. Evaluate how AI performs on complex matters involving ambiguity, competing authorities, evolving regulations, and high-stakes legal decisions.
Each project is different, but legal experts typically contribute in four ways:
Review and compare model outputs across realistic legal matters including contract analysis, litigation strategy, regulatory interpretation, and due diligence. Determine whether conclusions are legally supportable and consistent with factual record.
Develop gold-standard legal analyses, contract interpretations, and evaluation frameworks.. Your interpretation helps establish the standards of legal reasoning that frontier models are evaluated against.
Build legal scenarios where the correct answer is unclear. Assess how models perform when facts are incomplete, regulatory guidance is unsettled, or multiple legally defensible interpretations exist.
Apply expert oversight to identify flawed case interpretations, unsupported assumptions, and material legal risks. Ensure outputs meet standards for accuracy, defensibility, and reasoning quality.
Litigation Attorney Expert
Corporate & Compliance Attorney Expert
Specialist Legal Reviewer
We work with attorneys who have spent years navigating complex legal frameworks and high-stakes decisions to help evaluate and improve today's most advanced AI systems.
Evaluate how AI models read pleadings, opinions, discovery and the evidentiary records. Assess whether it can find the dispositive issue and tell controlling precedent from persuasive authority.
Assess whether how a model interprets contract language, allocates risk, and analyzes transaction documents against the f precision expected in sophisticated transactions demands.
Examine how AI approaches regulatory obligations, subpoenas, internal investigations, and enforcement actions. Determine whether recommendations account for evolving regulatory frameworks, disclosure requirements, and the practical realities of managing legal and compliance risk.
Assess model performance in legal domains where deep subject-matter expertise is essential. Identify whether analyses reflect genuine understanding of complex doctrines, industry-specific regulations, and specialized legal frameworks.
Evaluate how AI approaches workplace disputes, employee investigations, disciplinary actions, and employment agreements. Determine whether recommendations appropriately balance legal obligations, litigation risk, employee relations, and business objectives.
Strong candidates typically have professional legal experience and can independently reason through complex legal issues and factual scenarios.
Experience in litigation, transactional law, corporate law, compliance and regulatory practice, real estate, tax, bankruptcy, trusts & estates, legal research, or judicial clerkships is highly relevant for most roles.
No. The work is substantially closer to legal analysis, legal writing, issue spotting, and professional review than traditional prompt engineering.
Strong candidates typically have:
Compensation varies by project, specialization, and experience level.
Recent opportunities have included:
Qualified specialists and highly experienced legal professionals may receive access to additional premium opportunities.