Guardrails for lawyers

When you finish, you can apply confidentiality, privilege, competence and verification duties to your use of AI.

Confidentiality and privilege

Never put client-identifying or privileged information into a consumer AI tool unless your firm or employer has approved that tool for that data. Terms differ. Some tools may keep inputs, let staff review them or use them for training, depending on the plan and settings. A careless paste can create a confidentiality problem or argument about waived privilege.

Learn what your approved tool does with inputs, where data is stored, who can access it and whether it crosses borders. Data protection rules such as the GDPR in Europe or similar laws elsewhere may apply to personal data in documents. Ask your firm's compliance or technology lead rather than guessing.

Competence, supervision and the duty to verify

Professional rules in many places require lawyers to be competent in the tools they use and to supervise others, including non-lawyer assistants and systems. Bar associations and courts in many countries have published guidance on AI. Read what applies to you. This course is not legal advice and does not replace those rules.

Verification is the heart of it. AI tools can produce cases that do not exist, quotes that no judge wrote, and statutes from the wrong place. The fix is not to avoid AI but to treat every authority as unverified until you have read it at the source. Build the habit into your workflow, not your good intentions.

Bias, fairness and your client's story

AI summaries can flatten or tilt facts. In family, criminal, employment and immigration matters, a loaded description of a person can do real harm. Compare any summary to the original and watch for adjectives and assumptions that were not in the record.

Also be careful with tools that evaluate people, such as ranking witnesses or predicting credibility. These can embed biases from their training data. Use them, if at all, only with approval and clear understanding, and never as the decision.

Disclosure, billing and knowing when to refuse

Check court rules, engagement letters and client requirements on AI disclosure. Some courts require a statement or certification about AI use in filings. Be honest with clients about how you use these tools, and think about whether your billing reflects the real time spent.

Say no when you cannot verify the output, when the data is too sensitive for the available tool, or when asked to let AI give advice directly to a client. A short written AI policy for your team, even one page, removes much guesswork.

Practice now

Write a one-page AI use policy for yourself: approved tools, banned data, verification steps, disclosure notes.
Look up your own bar association's or regulator's published guidance on AI, and note two points that affect your work.

Use any AI assistant you have. Do not paste private, client or patient information into it.

Check yourself

1. What is the safest assumption about a free consumer chatbot and client data?

2. How should you treat a case citation from AI?

3. Why compare an AI summary of a client's story to the original?

Ask the AI tutor

Ask anything about this lesson, in any language. The tutor is an AI: it can be wrong, and it never needs your personal details.

Next lesson: Grow with AI: your 30-day plan

AI-drafted. Checked by an automated validator and a separate AI reviewer; a person has not read every lesson. Translations and tutor answers are AI-made. Found a mistake? Tell us. AI can be wrong. General information only, not professional advice. Not affiliated with any AI company named. Current University is not accredited.