Guiding Principles for Responsible Use of ChatGPT
- Be specific in prompts, including document type, audience, and purpose.
- Always maintain human oversight and independent review.
- Avoid entering confidential or protected information into non-enterprise systems.
- Never rely on ChatGPT for case law validation or citation accuracy.
- Document AI-assisted workflows when appropriate, particularly in regulated or client-sensitive matters.
- Disclose AI assistance when required by rules or client expectations.
- Anchor usage in ABA Model Rules: 1.1, 1.6, and 5.3.
Why Lawyers Use ChatGPT
Most legal work is information-dense, repetitive, and time-sensitive. ChatGPT can help reduce time spent on drafting, summarizing, outlining, and preparing for meetings, but only when used with proper human oversight. Clear prompts and thoughtful instructions dramatically improve the usefulness of its output.
Drafting Legal Documents with ChatGPT
Effective Prompting for Drafting
ChatGPT can generate strong first drafts when lawyers provide clear prompts identifying the type of document, purpose, audience, and key issues. Lawyers remain responsible for the accuracy, legal sufficiency, and appropriateness of the final document, including compliance with jurisdictional and ethical requirements.
Example Prompt
“Draft a first-cut engagement letter for a Pennsylvania estate-planning practice. Include sections on scope, fees, client responsibilities, and communication methods. Write in clear, client-friendly language.”
Practical Drafting Tips
- State the document type, purpose, and intended reader.
- List required clauses.
- Request tone adjustments or alternative versions.
Using ChatGPT to Summarize Legal Documents
ChatGPT can summarize large blocks of text so lawyers can focus on key points. This is particularly useful when preparing for meetings, depositions, or hearings. Lawyers must consider confidentiality obligations and firm policies when using AI.
Improving Client Communication
ChatGPT can rewrite complex legal language into clear, client-friendly English without altering meaning. This is useful for client updates, onboarding materials, and explanatory sections of agreements.
Organizing Information and Structuring Arguments
ChatGPT can convert rough notes into structured outlines, reorganize scattered ideas, and propose logical argument flow. This is useful for briefs, memos, presentations, and internal analyses.
Preparing for Hearings, Meetings, and Discussions
ChatGPT can draft talking points, timelines, key issues, and questions that may arise when provided with sufficient context.
Understanding the Limits of ChatGPT in Legal Research
ChatGPT cannot Shepardize or KeyCite, confirm whether a case is good law, or reliably summarize legal authority. It may generate citations that appear plausible but do not exist, are misattributed, or are outdated. All citations must be independently verified.
Examples of Legal Workflows Where ChatGPT Adds Value
- Preparing a first-cut engagement letter or client agreement
- Drafting client explainer emails and status updates
- Creating meeting agendas and talking points
- Turning deposition notes into a structured outline
- Summarizing lengthy regulatory guidance or policy documents
- Creating multiple tone-adjusted versions of the same message
When ChatGPT Should Not Be Used
- Highly confidential matters without secure enterprise controls
- Jurisdiction-specific legal analysis without independent verification
- Situations where court-mandated disclosure of AI use would create strategic or procedural risk
- Substantive legal analysis without primary source review
An increasing number of courts have issued standing orders requiring disclosure of generative AI use in filings, and lawyers have already been sanctioned for submitting AI-generated content containing hallucinated citations or factual errors. Courts have made clear that inadvertent reliance on AI hallucinations is not a defense.
Collaboration and Document Review
ChatGPT can summarize changes between versions, describe edits, and help merge writing styles across contributors.
Confidentiality and Data Handling Considerations
Lawyers must understand how AI vendors handle data, including retention, training use, and access controls. Consumer-grade tools may store or process data in ways that are incompatible with professional confidentiality obligations. Firms should evaluate whether enterprise versions or secure environments are required before permitting AI use in client matters.
Many firms are also exploring legal-specific or enterprise AI tools that offer features such as private instances, restricted data retention, and integrated citation verification. These platforms are often designed with professional confidentiality and regulatory requirements in mind and may be more appropriate for client work than general consumer AI tools. Lawyers should evaluate whether such solutions are necessary based on their practice areas, risk profile, and client expectations.
Governance, Disclosure, and Professional Responsibility
Lawyers must follow firm AI policies, ensure confidentiality, and maintain human oversight when using generative AI tools. Organizations should conduct vendor due diligence, understand data handling and metadata risks, and implement clear internal AI use policies. ABA Formal Opinion 512 (2024) reinforces many of these obligations, emphasizing competence, confidentiality, supervision, and the need to understand how AI tools function before relying on them in client work.
What Firm AI Policies Should Address
- Which tools are approved for use
- What types of matters or data are prohibited
- Training and supervision expectations
- Confidentiality and data handling requirements
- Disclosure obligations
Disclosure Obligations
Disclosure obligations may arise from court standing orders, client agreements, firm policy, or professional responsibility rules. Lawyers should review all applicable requirements before using generative AI in client work and should not assume that silence means permission. When using AI tools to draft or summarize documents, lawyers should consider disclosing such use to clients or courts, especially if required by local rules or client expectations.
Sample Disclosure Language
For example: “This document was prepared with the assistance of generative AI and reviewed by an attorney for accuracy and completeness.”
Common Pitfalls and How to Avoid Them
- Hallucinations: Always verify output against authoritative sources.
- Bias: AI tools may reflect or amplify biases present in their training data. Lawyers should review all output for fairness, accuracy, and unintended assumptions.
- Stay informed about disciplinary cases and court sanctions arising from improper AI use.
Quick Checklist Before Using ChatGPT
- Is the platform enterprise-grade and compliant with firm policies?
- Have you removed all confidential or identifying information?
- Will you independently review and verify all AI-generated content?
- Is disclosure to clients or courts required?
Conclusion and Next Steps
ChatGPT can meaningfully enhance drafting, summarizing, and communication, but it cannot replace legal judgment or jurisdiction-specific expertise. Thoughtful usage can improve efficiency, but only with proper oversight and professional accountability. Lawyers should stay updated on evolving ABA and state guidance and consider CLEs or firm trainings on AI to ensure responsible use.
This article was originally published on March 5, 2026, in the ABA’s Law Technology Today.