ChamberOne

AI skills on a lawyer's CV and in hiring decisions

By the ChamberOne team 4 minute read

Two conversations are going wrong at the moment. Candidates are adding 'AI proficient' to their CVs without saying what they can do. Hiring partners are asking 'do you use AI?' and taking a confident yes as evidence of skill. Neither tells anyone much. This article suggests what credible competence looks like, how to show it, and how to test it fairly.

Key points

  • Replace the label 'AI skills' with specific tasks and how you checked the output.
  • Candidates should be ready to describe a real mistake and how they caught it.
  • Hiring partners should ask for examples and test source-checking, not tool use.
  • Do not penalise honest inexperience, and keep core legal skills in view.

What competence actually is

Using AI well in legal work is less about the tool and more about three habits. Knowing which tasks suit it and which do not. Giving clear instructions and refining them. Checking the output against the source, every time.

A lawyer who has these habits can pick up a new tool in a week. A lawyer who lacks them can use the best tool badly. So the signal to look for, and to offer, is judgement.

For candidates: writing it on the CV

Avoid the bare label. 'Proficient in AI tools' says nothing and invites scepticism. Prefer a specific, modest statement of what you did.

  • Weak: 'Advanced AI skills.' Stronger: 'Used an AI assistant to produce first-draft issue lists for contract reviews; checked every item against the source document.'
  • Weak: 'AI-driven research.' Stronger: 'Used AI to suggest search terms for case law research, then ran the searches in the official databases and read the judgments.'
  • Weak: 'Prompt engineering expert.' Stronger: 'Wrote a short internal guide on instructing AI for summarising witness statements, including what to check.'
Weak
  • 'Advanced AI skills'
  • 'AI-driven research'
  • 'Prompt engineering expert'
Stronger
  • First-draft issue lists for contract reviews, every item checked against the source
  • AI-suggested search terms, then searches run in official databases and judgments read
  • A short internal guide on instructing AI for summarising witness statements
Writing AI skills on a CV

For candidates: in the interview

Expect questions about mistakes. A good answer names a real one: a case citation that did not exist, a summary that missed a qualification, a clause misread. Then it says how you caught it and what you changed. An answer in which nothing ever went wrong sounds either inexperienced or untruthful.

Be ready to discuss confidentiality. Say what you would never paste into a public tool, and how you would find out the firm's rules. Mention that you would check the guidance from the regulator or bar and the firm's own policy before using any tool on client material.

If you have not used AI at work, say so and describe how you have learned about it. Honest and curious is better than inflated.

For hiring partners: questions that test judgement

Replace 'do you use AI?' with questions that need a real example.

  • Tell me about a task where you decided not to use AI. Why?
  • Describe a time AI output was wrong. How did you notice?
  • How would you check an AI-generated summary of a long contract before relying on it?
  • What would you not put into a public AI tool, and why?
  • How would you explain to a client that AI was used on their matter?

A small practical exercise

For shortlisted candidates, consider a short, paid or clearly time-limited task. Give an anonymised, invented document and an AI-generated summary containing two or three planted errors, one of them subtle. Ask the candidate to review the summary and mark what is wrong.

You are not testing whether they can use a tool. You are testing whether they read the source. Candidates who find the subtle error are showing the skill that matters. Keep the exercise brief, fair and the same for everyone.

Do not use real client documents. If the task is more than a short exercise, pay for the candidate's time.

Do
  • Use an anonymised, invented document with planted errors
  • Test whether the candidate reads the source
  • Keep it brief, fair and the same for everyone
  • Pay for the candidate's time if it is more than a short exercise
Don't
  • Use real client documents
  • Treat it as a test of tool use
Setting a fair review exercise

Fairness and overreach

Do not penalise candidates for honestly saying they have limited experience, particularly those who trained or worked where tools were unavailable or restricted. Access to AI has been uneven across firms and countries, and a lack of experience is not a lack of ability.

Equally, do not require AI skills as a condition for roles where they are unnecessary, or let a candidate's enthusiasm for tools stand in for legal ability. Core skills such as analysis, drafting, advocacy and client care still decide the quality of a lawyer's work.

Hiring decisions that use AI screening tools bring their own concerns about bias and transparency. Check the guidance from your regulator and any employment authority before relying on automated screening, and keep a person accountable for each decision.

After the hire

Competence at interview is a starting point. Once someone joins, give them the firm's policy, an approved list of tools and a named person to ask. Pair new starters with a colleague for the first few AI-assisted tasks. Review early work with attention to checking habits, not just results.

Over time, treat AI use as part of ordinary supervision, not a separate speciality. The lawyers who thrive will be those who use tools with care and can explain what they did.

Describing learning, not only doing

Candidates who have had little chance to use tools at work can still show relevant effort. Describe a course taken, a personal experiment with a clearly harmless text, or a team discussion you led on how the firm might use tools safely. Say what you learned, including what surprised you.

Hiring partners should treat this as genuine evidence. A candidate who has thought carefully about risks and limits, even without a long record of use, will usually adapt faster than one who has used a tool often but never questioned it.

A note on honesty on both sides

Inflated claims travel in both directions. Firms advertise 'AI-enabled practice' while most staff still copy text into a browser tab. Candidates answer in kind. If a firm wants candid answers, it should describe its own use truthfully, including what is not yet in place.

A candidate who hears an honest description of the firm's tools and policy can say what they would bring and what they would need to learn. That is a better basis for a hire than two parties matching each other's exaggeration.

This article is general information about the law, legal practice and technology, current when it was published. It is not legal advice. Laws and guidance change: check the sources linked above and take advice on your own circumstances from a lawyer qualified where you are.

Read next