Legal Expert

Models produce fluent legal analysis that is often subtly wrong: invented citations, misapplied precedent, and jurisdictional confusion. A plausible wrong answer in law has real consequences.

As a legal expert you'll grade AI legal reasoning against the standard of a careful practitioner and write the analysis the model should have produced.

• Draft prompts that test how models handle contracts, statutes, procedure, and case analysis.
• Review AI output for invented citations, misapplied precedent, and jurisdictional errors, and rate the reasoning.
• Write the correct analysis when the model falls short, with the clarity of a well-drafted memo.