When AI Changes the Representation, What Does the Client Have a Right to Know?
ABA Model Rule 1.4, confidentiality, informed consent, and the professional responsibility to preserve meaningful client participation in AI-assisted legal representation.
Analysis by
Founder, Lex Wire Journal • Technology, Governance & Sovereignty Strategist
Consider a law firm handling a litigation matter involving thousands of confidential documents.
The attorney identifies an artificial intelligence platform that could accelerate review, organize evidence, identify relevant information, and assist in evaluating potential litigation strategies.
The technology appears useful. The lawyer intends to supervise the work and independently review any consequential recommendations.
But the platform also processes information through systems operated by a third-party provider. Its contractual terms govern access, retention, and possible further uses of the material.
The client has not been told about the proposed arrangement.
Must the lawyer explain the technology before using it? Does the client need to consent? Would the answer change if the system operated in a properly controlled environment with substantially different information handling protections?
And what if the AI tool were used not merely to organize documents but to recommend a settlement strategy?
These questions illustrate a central professional responsibility challenge emerging as AI becomes integrated into legal practice.
The ethical obligation is not simply to disclose every technology used by a lawyer.
It is to determine what information the client needs, when communication is required, when informed consent must be obtained, and how the client can meaningfully participate in decisions affecting the representation.
In my continuing legal education program, Client Communication in the Age of AI: Building Trust Through Transparency, Ethics, and Oversight , I examine these questions through the lens of ABA Model Rule 1.4.
The central argument is that artificial intelligence changes the circumstances in which lawyers must communicate, but does not replace the underlying professional duty.
“The duty to communicate has always required lawyers to explain matters that materially affect the representation. AI simply creates new situations where that explanation becomes necessary.”
Jeff Howell, Esq. | Client Communication in the Age of AI, CLE Instructional Materials
The distinction between disclosure and informed communication matters.
A lawyer may inform a client that AI was used without explaining anything material about its role. Conversely, a lawyer may provide a thorough explanation of the relevant analysis and risks even when the precise technology used is not itself a required subject of disclosure.
The objective is to ensure that technological efficiency does not undermine the client’s ability to make informed decisions.
The Bottom Line
Lawyers are not universally required to disclose every use of artificial intelligence. ABA Model Rule 1.4 requires appropriate consultation, information, and explanation so clients can make informed decisions about their representations. Separate duties, particularly Rule 1.6, may require informed consent when an AI system creates material risks of disclosing protected client information.
Responsible client communication therefore requires three distinct inquiries: whether disclosure or consultation is necessary, whether informed consent is required, and what the client must understand to participate meaningfully in the relevant decision.
Rule 1.4: Client Communication as the Foundation of Informed Representation
Discussions of artificial intelligence and legal ethics frequently begin with confidentiality or technology competence.
Both are essential. But communication deserves equal attention because clients cannot meaningfully participate in decisions about their representations without sufficient information.
ABA Model Rule 1.4 establishes several distinct communication obligations.
Under Rule 1.4(a)(1), lawyers must promptly inform clients about decisions or circumstances requiring informed consent under the applicable rules.
Rule 1.4(a)(2) requires reasonable consultation about the means by which the client’s objectives are to be accomplished.
Rule 1.4(a)(3) requires lawyers to keep clients reasonably informed about the status of their matters.
Rule 1.4(a)(4) addresses reasonable requests for information.
Rule 1.4(b) requires explanation sufficient to permit informed decisions concerning the representation.
These obligations are not technology-specific.
They concern how attorneys maintain a professional relationship in which clients understand material developments and retain appropriate authority over their objectives and decisions.
Not every choice of software requires client participation.
Lawyers regularly select research databases, document management systems, communication applications, and other tools without seeking approval for every routine decision.
But technology choices may become material when they affect confidentiality, significant professional decisions, client instructions, or the terms of the representation.
The relevant inquiry is not simply whether artificial intelligence is present.
It is whether the particular use creates circumstances that require consultation, explanation, or consent.
This is the distinction between a technology disclosure policy and meaningful professional communication.
Why Rule 1.4 Cannot Be Considered in Isolation
A lawyer’s communication obligations do not exist independently of other professional duties.
AI-assisted representation may involve competence, confidentiality, supervision, privilege, and fees.
Each can affect what the lawyer must understand before explaining a proposed technology use to the client.
Competence under Rule 1.1 matters because an attorney cannot accurately explain a system’s material risks without understanding its relevant capabilities and limitations.
Confidentiality under Rule 1.6 matters because information relating to a representation may be processed, stored, or accessed through AI infrastructure.
Supervision under Rules 5.1 and 5.3 matters because firm personnel must understand the boundaries of authorized use and accurately implement the firm’s commitments to clients.
Reasonable fees under Rule 1.5 matter because AI-assisted efficiency can affect the time and expenses associated with legal work.
Attorney-client privilege raises separate issues concerning protected communications and the consequences of disclosures to third parties.
The ethical duty of confidentiality is broader than the evidentiary attorney-client privilege.
The involvement of a third-party service does not automatically resolve whether privilege has been waived.
That question depends on the governing law, the nature of the information and disclosure, and the circumstances surrounding its protection.
The lawyer must not confuse these separate legal principles when counseling the client about AI-related risks.
This broader structure explains why the same AI system can present different communication obligations depending on the work being performed.
Three Distinct Questions: Disclosure, Informed Consent, and Client Decision-Making
The terms disclosure, consent, and transparency are often used interchangeably in conversations about AI ethics.
That approach obscures important distinctions.
A lawyer may be required to communicate information without needing client consent for every related decision.
In other circumstances, a lawyer must obtain informed consent before proceeding.
And even when a formal AI disclosure is unnecessary, the lawyer may still need to explain material risks, alternatives, and recommendations so the client can make an informed decision.
Determine whether the lawyer must explain the proposed AI use or consult with the client under Rule 1.4, an agreement, or another applicable obligation.
Determine whether a separate professional rule, including Rule 1.6, requires the client’s informed agreement before the proposed technology use proceeds.
Explain material options, risks, limitations, assumptions, and consequences so the client can meaningfully participate in representation decisions.
This three-part distinction is a proposed analytical framework for applying existing professional duties.
It is not an additional ABA ethics rule or a mandatory three-step procedure for every AI use.
Its purpose is to ensure that lawyers identify which obligation is actually involved.
Consider a lawyer using a general-purpose AI tool to generate a list of broad legal research topics without entering information relating to a representation.
That activity may create no independent AI-specific duty of client disclosure or informed consent, absent additional relevant circumstances.
Now consider a lawyer proposing to transmit sensitive client information into a self-learning AI platform whose outputs could disclose that information to unauthorized users.
The confidentiality analysis becomes materially different.
Finally, consider a lawyer relying on AI-assisted predictions when recommending settlement.
The client’s need to understand the analysis and its limitations may raise Rule 1.4 concerns even where no separate information disclosure risk exists.
These situations should not be collapsed into a single rule that every AI interaction requires the same communication response.
The No Surprise Rule: Recognizing Material AI Use
In my CLE program, I introduce a practical concept I call the No Surprise Rule.
It begins with a simple question:
Would a reasonable client be materially surprised to learn how their information was used, how the legal work was performed, or how a consequential recommendation was developed?
If the answer is yes, the attorney should examine whether communication or informed consent is required.
“Transparency builds trust. Silence creates surprise.”
Jeff Howell, Esq. | Client Communication in the Age of AI, CLE Instructional Materials
The No Surprise Rule is not the legal standard established by Model Rule 1.4.
A client can be surprised by an entirely permissible technology.
A client might also expect widespread AI use in circumstances where separate confidentiality obligations nevertheless require informed consent.
The value of the heuristic is that it prompts the lawyer to investigate potentially material circumstances before proceeding.
Several situations deserve particular attention.
Changes in Information Handling
The proposed AI workflow may transmit client information to an external provider or materially change access, retention, or reuse arrangements.
The lawyer should evaluate whether the system is appropriate and whether consent or communication is required.
Consequential AI Recommendations
A system may materially influence a settlement assessment, case strategy, or other significant decision.
The client may need to understand the method’s limitations and the lawyer’s independent evaluation.
Significant Cost or Scope Changes
AI-assisted efficiency may affect billing expectations, agreed procedures, or the scope of work.
The relevant fee agreement and professional obligations determine what must be communicated.
Client Instructions and Expectations
A client may impose restrictions on the use of AI or require notification under engagement terms or outside counsel guidelines.
Those instructions must be considered before the lawyer chooses the workflow.
The practical lesson is that silence should not be mistaken for agreement.
A lawyer cannot infer that a client has received necessary information merely because the client has not asked about artificial intelligence.
Does a Lawyer Have to Disclose the Use of AI to a Client?
Not automatically.
Whether disclosure is required depends on the applicable professional rules, the nature of the AI use, client instructions, contractual obligations, and the circumstances of the representation.
ABA Formal Opinion 512 , issued July 29, 2024, provides the central ABA guidance for generative AI.
The opinion separates the Rule 1.6 confidentiality analysis from the Rule 1.4 communication analysis.
It recognizes that disclosure may be unnecessary in some circumstances.
It also identifies situations where lawyers must consult with their clients about AI use.
For example, the opinion addresses client inquiries about the use of generative AI, express client instructions, fee-related issues, and AI outputs influencing significant decisions in a representation.
The opinion’s confidentiality discussion is particularly important.
It explains that because certain self-learning generative AI tools may disclose information relating to a representation to unauthorized users, informed client consent is required before entering such information into tools presenting the described risks.
That conclusion should not be generalized into a claim that all AI tools present identical information handling risks.
Different systems, configurations, contractual arrangements, and proposed uses require separate evaluation.
| Illustrative AI Use | Communication Consideration |
|---|---|
| General brainstorming without client information | An AI-specific disclosure or consent duty may not arise, absent other material circumstances. |
| Routine administrative assistance | May not require specific disclosure if it does not materially affect the representation or trigger another obligation. |
| AI-assisted drafting with appropriate lawyer review | Context-dependent. Routine assistance is not automatically subject to client consent. |
| Sensitive client information entered into a system with disclosure risks | Evaluate Rule 1.6 and whether informed consent is required before proceeding. |
| AI-assisted litigation prediction influencing significant advice | Consider Rule 1.4 consultation and the explanation needed for an informed client decision. |
| Engagement agreement restricts AI use | Evaluate and honor the relevant client instructions and contractual obligations. |
The table illustrates the need for a fact-specific assessment.
It is not a categorical classification of every possible AI application.
The same platform may support a low-risk general administrative task and a sensitive client-specific analysis.
The relevant duties must be evaluated for the actual use.
The AI Communication Decision Framework: Five Questions Before Proceeding
Legal standards are more useful when attorneys can apply them through repeatable decisions.
My CLE includes a decision flowchart for identifying when AI use raises client communication concerns.
The following adaptation preserves that practical approach while distinguishing legal requirements from recommended professional practices.
The Lex Wire AI Communication Decision Framework
-
Identify the
Proposed AI Use
What technology will be used, for what task, and how will it affect the representation?
-
Trace the
Information Flow
Will the application process information relating to the client? Who can access, retain, or reuse that information?
-
Assess Material
Consequences
Could the use affect confidentiality, privilege, legal strategy, fees, scope of work, or important client decisions?
-
Determine the
Applicable Duty
Does Rule 1.6 require informed consent? Does Rule 1.4 require consultation or explanation? Do the engagement terms impose additional conditions?
-
Communicate and
Document Appropriately
Explain relevant benefits, risks, safeguards, and alternatives. Obtain consent when required and record material instructions or authorizations.
The decision framework separates two activities that should not be confused.
The first is understanding what the technology actually does in the representation.
The second is determining what legal and professional obligations follow from that use.
A firm cannot reliably determine its communication responsibilities without understanding the workflow, the information involved, and the relevant vendor practices.
This is particularly important where attorneys rely on representations made by software providers.
Terms such as private, secure, or enterprise-grade do not, standing alone, establish the actual information protections relevant to client consent.
How to Explain AI Use to Clients: The Four-Part Informed Consent Framework
Once a lawyer determines that a client should be consulted, the next challenge is communicating clearly.
Clients generally do not need a technical lecture about neural networks, training methods, or model architecture.
They need accurate information about the proposed use, its material benefits, relevant risks, and the safeguards available to protect their interests.
In my CLE, I present a four-part Informed Consent Conversation Framework .
Although originally developed for consent-related discussions, the same communication structure can assist lawyers in explaining AI practices when formal consent is not required.
Explain why the lawyer proposes using AI and what the technology may contribute to the client’s objectives.
Explain material limitations, information handling concerns, uncertainty, and potential consequences.
Describe actual protections, approved workflows, contractual controls, and meaningful attorney supervision.
Encourage questions, explain relevant alternatives, and obtain informed consent where required.
Lead With Value
A useful conversation begins by explaining why the lawyer believes AI may assist the representation.
For example, an approved document review platform may help organize a large collection of evidence or identify material requiring further attorney evaluation.
A lawyer should explain those potential advantages in terms the client can understand.
But the explanation should not promise faster results, lower costs, or greater accuracy unless those claims have a reasonable factual basis.
Acknowledge Risk
The lawyer should identify material risks associated with the particular use.
Depending on the system, those risks may include inaccurate outputs, retention of client information, possible unauthorized access, or limitations in the system’s analysis.
The purpose is not to frighten the client.
It is to support an accurate understanding of the proposed arrangement.
Risks should neither be exaggerated nor minimized merely to obtain agreement.
Explain Safeguards
Clients should understand the relevant steps the firm takes to protect information and review AI-assisted work.
These may include approved software, access controls, contractual confidentiality provisions, retention restrictions, and attorney verification.
The critical requirement is accuracy.
A lawyer should not tell a client that information will never be retained or used for training without verifying the applicable terms and technical arrangements.
Nor should the firm describe human review as comprehensive if its actual procedure consists only of superficial approval.
Invite Client Decision-Making
The final step is to make the conversation interactive.
Clients should have an opportunity to raise relevant concerns, ask questions, and understand available alternatives.
If informed consent is required, the attorney must provide the information necessary for an informed agreement under the applicable rules.
Where consent is not required, communication can still be appropriate to protect the relationship and clarify expectations.
“Rule 1.4 is not satisfied by awareness. It is satisfied by informed decision making.”
Jeff Howell, Esq. | Client Communication in the Age of AI, CLE Instructional Materials
This statement captures the professional purpose behind the framework.
A standardized conversation can improve consistency, but it does not automatically satisfy Rule 1.4 or Rule 1.6.
The content must be sufficient for the particular circumstances and legal obligations involved.
Three Practical AI Client Communication Scenarios
The following hypothetical scenarios demonstrate how the communication analysis changes depending on the proposed use of AI.
They are not representations of documented incidents involving particular law firms or vendors.
External AI Document Review and Confidential Client Information
A firm proposes using an external AI platform to organize and summarize confidential discovery materials.
The vendor processes uploaded documents through its own infrastructure, subject to specific retention, access, and contractual arrangements.
Before proceeding, the attorney must evaluate whether the proposed processing is compatible with confidentiality obligations and client instructions.
This scenario requires more than checking whether the tool can accurately summarize documents.
The lawyer should understand who can access the information, whether it can be reused, how long it is retained, and whether contractual protections apply.
If the system presents the disclosure risks described in ABA Formal Opinion 512, informed consent may be required before the information is entered.
The lawyer should explain the proposed use, relevant benefits, material risks, available safeguards, and any alternatives necessary for an informed decision.
If the firm instead uses an approved system with materially different controls, the analysis may change.
The important distinction is not whether the provider markets the software as legal AI.
It is the actual information handling arrangement and the duties applicable to the representation.
AI-Assisted Settlement Evaluation
A lawyer uses AI to organize past outcomes, compare factual patterns, and prepare a preliminary assessment of litigation risk.
The system recommends a settlement range that the attorney believes merits serious consideration.
The client must decide whether to accept an offer, continue negotiating, or proceed toward trial.
Here the communication issue is not limited to explaining that an AI tool was used.
The client needs to understand the material basis of the lawyer’s recommendation.
That may include the evidence, competing scenarios, uncertainty, and limitations affecting the analysis.
ABA Formal Opinion 512 specifically identifies significant decisions, including AI-assisted litigation outcome evaluations, as circumstances that may require client consultation.
The lawyer should also explain the independent professional evaluation applied to the AI-generated material.
Clients should not be left with the impression that a predictive output represents a guaranteed or objectively determined legal outcome.
The professional responsibility is to counsel the client about the available options, not merely relay the machine’s prediction.
This connects directly to The Ethics of Artificial Intelligence in Legal Practice: Professional Judgment in the Age of Automation , which examines the lawyer’s responsibility to independently evaluate AI-assisted recommendations.
AI Efficiency, Billing, and Client Expectations
A firm uses an approved AI application to assist with a substantial drafting assignment.
The work requires fewer attorney hours than the firm originally anticipated.
The client asks whether the efficiency affects the amount billed.
The lawyer should first consider the agreed fee arrangement and Rule 1.5’s reasonableness requirements.
For hourly billing, lawyers cannot charge for time they did not actually spend.
Flat fees and other arrangements require analysis under the applicable rules and client agreement.
The firm should not assume that every efficiency improvement automatically requires a revised pricing model.
Nor should it represent time savings as attorney work that was never performed.
Communication becomes particularly important when AI use affects the basis or reasonableness of a fee, or when the client seeks information about how the work was completed.
Formal Opinion 512 addresses both fee obligations and the potential communication implications of AI-assisted efficiency.
The broader professional question is how to explain the relationship between technology, actual attorney work, and the value of the service provided.
Engagement Letters, Documentation, and Institutional Consistency
Individual lawyers need sufficient information to communicate accurately with clients.
Law firms also need procedures that support those conversations consistently.
Engagement letters can be useful for describing the firm’s general approach to AI-assisted legal work.
They may explain the nature of authorized uses, the continuing responsibility of attorneys, and how clients can ask questions or communicate instructions.
However, general engagement language is not always sufficient when a particular use requires informed consent.
Formal Opinion 512 specifically cautions that general boilerplate is insufficient to secure informed consent in the confidentiality circumstances addressed by the opinion.
Firms should therefore distinguish general notification from matter-specific consultation and consent.
Material client instructions should be communicated internally to personnel responsible for the relevant work.
Consent records should be maintained where appropriate, with access and retention practices consistent with confidentiality.
The firm should also have a process for identifying material changes in vendor terms or technology functions that warrant renewed review.
These are governance functions, but their purpose includes supporting accurate client communication.
The broader governance structure is examined in Building an AI Governance Framework for Law Firms: Six Pillars of Operational Accountability .
For practical documentation procedures, see Building Your Law Firm’s AI Playbook .
Trust Is Built Through Informed Participation, Not Disclosure Alone
Law firms increasingly discuss transparency as a way to build trust around AI adoption.
The principle is sound, but disclosure by itself should not be mistaken for effective communication.
A lawyer can provide a technically accurate statement about the use of AI without helping the client understand anything material about the representation.
A long technology disclaimer may also overwhelm clients with information unrelated to their decisions.
Meaningful communication requires selecting the information that matters.
The attorney should explain material risks, relevant alternatives, and the professional basis of significant recommendations in language the client can understand.
This is especially important when the client faces consequential decisions involving litigation, business interests, financial exposure, personal circumstances, or sensitive information.
AI can help organize the information that informs those conversations.
It does not eliminate the lawyer’s responsibility to explain the relevant choices.
In my view, the strongest opportunity for AI-assisted practice is not simply completing existing tasks more quickly.
It is creating more room for attorneys to focus on analysis, counseling, and client understanding.
That benefit is not automatic.
It depends on how the firm uses technology and whether the time saved is redirected toward meaningful professional work.
The Bottom Line: AI Changes the Conversation, Not the Duty
Artificial intelligence is changing how legal information is processed, how documents are prepared, and how recommendations are developed.
But these changes do not eliminate the lawyer’s existing duty to communicate with clients.
Rule 1.4 remains the central framework for determining what clients need to understand in order to participate meaningfully in their representations.
Rule 1.6 provides separate obligations concerning confidential information, including circumstances requiring informed consent.
The professional challenge is applying those duties accurately as AI introduces new information flows, capabilities, and decision-making processes.
The No Surprise Rule helps identify potentially material issues.
The communication decision framework helps determine which duty applies.
The four-part conversation model helps attorneys explain relevant benefits, risks, safeguards, and alternatives when communication is appropriate or required.
“AI requires more communication, not less.”
Jeff Howell, Esq. | Client Communication in the Age of AI, CLE Closing Remarks
That statement expresses my professional perspective, not a universal mandate to disclose every AI-assisted activity.
The obligation is to communicate what the circumstances and applicable rules require, while recognizing that thoughtful transparency may also serve the client relationship.
The ethical objective is not merely to tell clients that AI was used. It is to preserve their ability to understand and participate in decisions that materially affect their representation.
Related Lex Wire Journal Analysis
- The Ethics of Artificial Intelligence in Legal Practice: Professional Judgment in the Age of Automation examines how attorneys preserve independent professional judgment as AI performs more legal work.
- Ethics in the Age of AI: Key Takeaways for Attorneys summarizes the original NACLE ethics presentation.
- Building an AI Governance Framework for Law Firms addresses institutional responsibility, risk classification, and operational AI oversight.
- Building Your Law Firm’s AI Playbook provides practical procedures for translating governance principles into firm operations.
- Who Controls Your Law Firm’s AI Intelligence Layer? examines institutional knowledge, governance, and technology control.
Legal Authorities and Source Materials
Professional Responsibility
- American Bar Association, Model Rule 1.4: Communications , addressing client consultation, information, and explanation.
- American Bar Association, Comments to Model Rule 1.4 .
- American Bar Association, Formal Opinion 512: Generative Artificial Intelligence Tools (July 29, 2024), particularly the confidentiality and communication discussions.
- American Bar Association, Model Rules of Professional Conduct , including Rules 1.0(e), 1.1, 1.4, 1.5, 1.6, 5.1, and 5.3.
Original CLE Instructional Materials
- Jeff Howell, Esq., Client Communication in the Age of AI: Building Trust Through Transparency, Ethics, and Oversight , Lawline CLE presentation outline, expanded script, and instructional notes.