AI Training Roles

Lawyer and Legal AI Training Interview Questions

Legal roles ask attorneys and counsel to draft and review legal work product, write hard legal questions and judge model answers on accuracy and reasoning. Many postings ask for a law degree and bar admission. These practice questions focus on precise analysis and citing the right authority.

66 open roles in this groupUpdated 9 October 2026

Practice questions

These are practice questions written for this site to help you prepare. No platform has said it asks these exact questions.

  1. Review this model-drafted contract clause. What risks does it create?

    A good answer covers: Ambiguities, missing definitions, one-sided terms, conflict with other clauses, and a redline that fixes them.

  2. Write a legal question that requires multi-step reasoning, with a model answer.

    A good answer covers: A fact pattern, the governing rule, application to facts and a conclusion, with jurisdiction stated.

  3. A model cites a case that may not exist. What do you do?

    A good answer covers: Verify the citation in a legal database; if it cannot be found, mark it as fabricated and score accuracy down.

  4. How do you explain the IRAC method?

    A good answer covers: Issue, rule, application, conclusion; mention how it structures a good answer.

  5. How would you rate an answer that is correct in one jurisdiction but wrong in another?

    A good answer covers: Check what jurisdiction the prompt assumes; penalise if the answer ignores it or fails to flag differences.

  6. Explain an indemnification clause to a small business owner.

    A good answer covers: Plain words, who pays for what loss, and why limits matter.

  7. What are the limits of an AI legal answer, and how should it say so?

    A good answer covers: It is not legal advice for a specific case; it should suggest consulting a licensed lawyer for decisions.

  8. Describe your practice area and where you would defer to others.

    A good answer covers: Honest scope of expertise and jurisdictions.

  9. How do you mark up a document so another lawyer understands your changes?

    A good answer covers: Tracked changes, short comments explaining each change, consistent defined terms.

What the platforms say about the assessment

In the 2026-10-09 data, roles in this group came from micro1, RWS TrainAI, SME Careers, Meridial, Welo Data (Welocalize), Terac and xAI. Each platform runs its own process, and steps can differ by role. Below is what each platform's public pages or postings say, followed by lines quoted from current postings in this group.

micro1

Source: www.micro1.ai

RWS TrainAI

Source: api.lever.co

SME Careers

Sources: sme.careers, api.sme.careers

Meridial

Source: boards-api.greenhouse.io

Welo Data (Welocalize)

Source: welodata.ai

Terac

Source: terac.com

xAI

Source: boards-api.greenhouse.io

Quoted from current postings

"During a remote voice interview, you will discuss your organization's approach to data compliance and risk management."

Healthcare Data Governance Leaders: Interview on AI Access and Compliance, terac

"Specific details and role specific information will be provided to you during the interview process."

AI Tutor - Legal & Compliance, xai

"Selection process involves an AI interview, a domain-specific task, and an interview with a recruiter."

Legal Quality Assurance Lead (QAL), sme

Steps were read from public pages and postings, not tested first-hand. Passing a screening does not guarantee a project, hours or pay.

Terms to know

Domain expert
Someone hired for professional knowledge in a field, such as law, medicine or engineering, to write or judge expert-level tasks.
Hallucination
When a model states something false or made up as if it were true, such as a fake citation or an invented fact.
Grounding
Whether an answer is supported by a given source, such as a document, search result or retrieved passage.
Rubric
A written list of criteria and scores used to judge a response, for example accuracy, instruction following and tone, each with clear pass or fail descriptions.
Fact-checking
Checking each claim in a model's answer against reliable sources and marking what is wrong or unsupported.

Open roles in this group (66)

Showing the 40 newest. Search all roles

Sources

Other role groups