Can Your Law Firm Use ChatGPT or Claude for Client Work?

What could change for your business?

The aim is to move from a vague question about AI to a documented, bounded business decision. Relevant potential outcomes include:

• Identify appropriate and restricted client-matter uses.
• Compare tool settings and current commercial protections.
• Create review procedures for research, citations and client communication.
• Record when local ethics advice or client consent may be required.

Why is this question different?

This subject merits a distinct answer because it addresses professional duties under applicable local conduct rules, not just corporate data privacy. The relevant evidence and business decision differ from nearby pages in these ways:

• Professional duties under applicable local conduct rules, not just corporate data privacy.
• Comparative ChatGPT/Claude workflow focused on client information and legal work.
• Matter-level supervision, citation validation and client communication decision.

What supports the guidance?

This guide relies on first-party provider documentation and published institutional guidance, identified below. It does not claim to have inspected your private contract, account settings or security controls. The same product name may represent different plans, features and legal relationships. Therefore, no case study, result, approval or guarantee should be inferred from this page.

Detailed research: questions, evidence and practical safeguards

Direct answer: Lawyers can consider AI tools for research and drafting, but the relevant decision concerns client confidentiality, competence, supervision and professional responsibilities. The selected product plan and matter-specific information both matter.

Scope: This analysis focuses on the described business decision and the cited primary materials. It is not a substitute for review of an executed agreement, a specific account configuration or professional advice under the governing law. Each subsection identifies a reason the answer may change and a document or test that would help make the decision defensible.

Start with the duty owed to the client

Legal work presents obligations of competence, confidentiality, communication, supervision and candor that do not disappear when a drafting tool is introduced. The professional duty attaches to the practitioner and matter; it cannot be outsourced to a model provider by adopting a commercial workspace.

Evidence to examine: Review the jurisdiction, lawyer role, intended task, client engagement obligations and any governing bar or court guidance before classifying a use case. Practical response: Write a matter-level decision rule stating which work may be assisted by AI and who remains responsible for the final legal judgment.

Classify client information before using a tool

Matter files may contain privileged communications, trade secrets, litigation strategy and personal data. Some uses may be inappropriate even when the tool offers enterprise privacy features. A firm should not infer permission to disclose a file solely from the vendor’s statement that the information is excluded from default model training.

Evidence to examine: Separate public pleadings, already-disclosed materials, client-confidential information and restricted case files; check engagement letters and client instructions. Practical response: Require a documented confidentiality and access assessment before uploading representative client records or providing broad repository access.

Relevant primary reference for this part of the review: ABA announcement of Formal Opinion 512

Compare the exact ChatGPT and Claude plans

Commercial services and consumer accounts may operate under different terms, feedback rules, retention arrangements and controls. Vendor-brand shorthand cannot resolve the comparison. A lawyer testing a personal account may be subject to rules different from those governing a firm-managed environment purchased under organizational terms.

Evidence to examine: Capture the tool, plan, administrator settings, relevant contract, and feature or connector used for each workflow, without assuming defaults remain unchanged. Practical response: Issue an approved-tool list by task and account type; explain why a particular environment is permitted, restricted or subject to additional checks.

Apply ABA Formal Opinion 512 carefully

ABA Formal Opinion 512 addresses the use of generative AI under existing professional responsibility duties. It is guidance from the ABA rather than a uniform replacement for every state rule. A federal or state tribunal, specialist practice or particular client contract can impose more specific obligations than general guidance summarizes.

Evidence to examine: Use the ABA opinion as a starting point, identify controlling jurisdictional ethics rules and check court-specific orders or professional notices where applicable. Practical response: Escalate uncertain questions about consent, confidentiality and representations to counsel qualified for the lawyer’s jurisdiction and matter.

Relevant primary reference for this part of the review: ABA guidance on ethics and generative AI

Verify research sources and citations

Generated quotations, statutory references and case names can appear plausible even when inaccurate or nonexistent. Legal filings demand independent verification. A defective citation can mislead a client or tribunal and create professional consequences; checking style alone is not equivalent to checking authority.

Evidence to examine: Open every material cited source, confirm the quoted text and procedural status, and review whether the law governs the relevant jurisdiction and date. Practical response: Use a citation-verification checklist with the name of the responsible reviewer and do not permit unsupported automated citations into filed work.

Control document drafting and redlining

AI assistance with contracts may accelerate first drafts, but the tool may omit negotiating history, legal exceptions, special definitions or confidential commercial instructions. The risks vary according to whether the document is for internal brainstorming, client circulation, regulatory filing or a binding transaction.

Evidence to examine: Review the client template, original instructions, version history, tool input, final redline and approval chain; distinguish suggestions from accepted provisions. Practical response: Require lawyer sign-off on risk-allocation clauses and accurate description of what has and has not been reviewed.

Relevant primary reference for this part of the review: OpenAI Services Agreement

Decide when clients should be informed

Client communication expectations differ by matter and applicable duties. Whether express informed consent is needed may depend on the information disclosed and technology involved. An automatic notice to every client can be unhelpful, while silence about a meaningful confidentiality exposure can be inappropriate.

Evidence to examine: Check the nature of the work, vendor processing, the client’s restrictions, legal jurisdiction and internal risk classification before determining a communication path. Practical response: Document the rationale, provide understandable explanations where required, and seek matter-specific guidance when duties are uncertain.

Supervise staff, contractors and AI-enabled processes

Support staff, associates and external vendors may interact with AI systems through different devices and accounts. Weak role controls can defeat a well-written policy. The firm needs a practical supervision model that covers training, review and mistakes without assuming every tool exposes the same audit controls.

Evidence to examine: List approved accounts and uses, client-matter access permissions, verification steps, supervision assignments and the escalation procedure for an improper upload. Practical response: Run a small controlled workflow review before expanding AI use across teams; reassess when new connectors or automation features are introduced.

Record an auditable matter-use decision

An actionable law-firm policy identifies permitted tasks, prohibited data categories, review standards, retention responsibilities and the person accountable. Different practices within one firm may require different controls, especially for litigation, regulated transactions or cross-border client work.

Evidence to examine: Keep the current provider terms and policies with a dated decision record; include the known limitations and links to the controlling professional materials. Practical response: Review the policy on a defined cadence and on material vendor changes, rather than treating one broad AI authorization as permanent.

Worked decision example

A law firm considers using an AI assistant to summarize a confidential witness interview and help prepare an outline. The firm needs to verify its professional duties and client restrictions, choose an approved account, and require a lawyer to verify every quoted fact and legal authority before the output is used.

The useful outcome is a written decision that identifies what information can be used, which supplier or professional rule applies, who approves the activity, and what evidence remains missing. The example does not establish that the hypothetical business has obtained a legal clearance, provider certification or binding professional approval.

Evidence checklist and owner matrix

What this analysis does not establish

No review of public documentation can prove the privacy posture of your particular account, determine client consent, ensure security, clear third-party intellectual-property rights or substitute for licensed local legal advice. The evidence needs to match the actual user, product, jurisdiction and contractual record. Technical penetration testing, professional certifications and vendor-endorsed approvals are not represented as completed services.

Primary sources and current product information

The following official sources provide starting points. Verify the current applicable version and any executed agreement before relying on a material proposition.

• ABA announcement of Formal Opinion 512
• ABA guidance on ethics and generative AI
• OpenAI Services Agreement
• Anthropic Legal Center — current commercial and service terms
• Anthropic commercial model-training policy