Prompt engineering — the skill of crafting effective instructions for AI language models — is becoming an essential competency for modern legal professionals. Teaching it to law students prepares them for a profession increasingly shaped by AI tools.
1. Why Law Students Need This Skill
AI tools like ChatGPT, Harvey, and CoCounsel are already used in legal practice for:
- Drafting contracts and correspondence
- Summarising case law and legislation
- Legal research and analysis
- Due diligence review
Students who can use these tools effectively and critically will have a significant advantage in the job market.
2. Core Prompt Engineering Principles for Law
| Principle | Example |
|---|---|
| Specify jurisdiction | "Under English law..." not just "What is the law on..." |
| Set a role | "Act as a senior solicitor advising a corporate client..." |
| Define the output format | "Provide your answer as a structured legal memo with headings" |
| Include constraints | "Cite only cases from 2010 onwards" or "Do not include US authorities" |
| Iterate and refine | "Now focus specifically on the duty of care element..." |
3. Teaching Activities
- Prompt comparison exercise: Give students the same legal question and ask them to write prompts of varying quality. Compare the AI outputs and discuss why better prompts produce better results
- Error detection: Show students AI-generated legal text containing errors (hallucinated cases, incorrect citations, outdated law). Ask them to identify and correct the errors
- Iterative refinement: Start with a basic prompt and progressively improve it through 3–4 iterations, showing how each refinement improves the output
- Ethical discussion: Debate when it is appropriate to use AI in legal work and what disclosure obligations exist
4. Critical Evaluation Skills
The most important lesson is that AI output must always be verified. Teach students to:
- Check every case citation against Westlaw or LexisNexis
- Verify that statutes cited are current and in force
- Assess whether the legal analysis is accurate and complete
- Identify potential biases in the AI's training data
5. Assessment Integration
Consider incorporating prompt engineering into assessments: "Using an AI tool, draft a client advice letter on [topic]. Submit both your prompts and the final output, with a reflective commentary explaining your prompt strategy and any corrections you made."