Issuing Body and Document Basis
The Florida Bar's Professional Ethics Committee issued guidance addressing attorney use of artificial intelligence tools under the Florida Rules of Professional Conduct. The 2024 opinion — formally designated as Ethics Opinion 24-1 — addresses how existing professional responsibility rules apply when a Florida-licensed attorney uses AI-assisted tools for legal work, including drafting, research, and document review.
The opinion does not create new rules. It interprets how Rules 4-1.1 (Competence), 4-1.6 (Confidentiality of Information), 4-5.1 and 4-5.3 (Supervision), and 4-1.5 (Fees) apply to AI-assisted workflows. Florida attorneys are expected to read this opinion alongside the underlying rules, not as a standalone compliance checklist.
Competence: What the Opinion Requires
Rule 4-1.1 requires attorneys to provide competent representation, which includes the legal knowledge, skill, thoroughness, and preparation reasonably necessary for the representation. The 2024 opinion extends this to AI tool use in two directions.
First, an attorney must understand the AI tool well enough to evaluate its output. This does not require technical expertise in machine learning, but it does require knowing what the tool does, what its failure modes are, and where its outputs require independent verification. An attorney who submits AI-generated research without checking citations, or who uses an AI drafting tool without reviewing the substance of what it produced, likely falls short of the competence standard.
Second, the opinion notes that competence includes staying current with changes in the law and practice — including the benefits and risks of relevant technology. This language tracks Comment [8] to ABA Model Rule 1.1, which Florida adopted in substance. The practical implication: attorneys cannot claim ignorance of AI-related risks as a defense to a competence complaint.
Confidentiality: The Core Restriction
Rule 4-1.6 prohibits revealing information relating to representation of a client without informed consent, except as permitted by the rules. The 2024 opinion applies this directly to AI tool use: before inputting client information into any AI system, the attorney must assess whether doing so constitutes a disclosure of confidential information.
The opinion identifies several questions attorneys should answer before using an AI tool with client data:
- Does the vendor retain, use, or share input data for model training or other purposes?
- Is the data processed on infrastructure controlled by the vendor, a third-party cloud provider, or both?
- Does the vendor's data processing agreement provide confidentiality protections consistent with Rule 4-1.6?
- Can the attorney verify that the vendor's stated data-retention policies are actually enforced?
- Has the client given informed consent to the use of AI tools that may process their information?
The opinion does not prohibit using AI tools with client data. It requires that attorneys take reasonable measures to prevent unauthorized disclosure — which, in practice, means conducting due diligence on the vendor's data handling before deployment, not after a breach.
Supervision Obligations
Rules 4-5.1 and 4-5.3 govern supervision of lawyers and non-lawyer assistants respectively. The 2024 opinion treats AI tools as falling within the supervision framework — not as autonomous actors, but as tools whose outputs must be reviewed by a responsible attorney.
The practical effect is that a partner or supervising attorney cannot delegate AI-assisted work to a junior associate and treat the AI's output as reviewed simply because a human touched the file. The responsible attorney must ensure that someone with sufficient competence has actually evaluated what the AI produced. This applies whether the AI is used for research, drafting, document review, or any other legal task.
The opinion also addresses law firm obligations at the firm level. A firm that deploys AI tools across its practice must have policies and procedures in place to ensure that AI-assisted work product meets the same professional standards as work produced without AI assistance.
Fees and Billing
Rule 4-1.5 requires that fees be reasonable. The opinion addresses a specific question that has come up frequently in practice: if AI reduces the time required to complete a task, can an attorney still bill the same amount of time it would have taken without AI?
The opinion's answer is nuanced. Billing for actual time spent reviewing and verifying AI output is permissible. Billing for time that was not actually spent — because the AI compressed the work — raises a reasonableness problem under Rule 4-1.5. The opinion does not prohibit billing for AI tool costs as a separate disbursement, but it notes that passing through AI costs requires transparency with the client, consistent with the general rule on billing for expenses.
Disclosure to Clients and Courts
The 2024 opinion does not impose a blanket requirement to disclose AI use to clients. It does require disclosure where the client's informed consent is needed — primarily in the confidentiality context, where client data will be processed by a third-party AI system.
On court disclosure, the opinion notes that separate court rules or standing orders may impose their own AI disclosure requirements, and those obligations exist independently of the ethics opinion. Florida attorneys practicing in federal courts with AI standing orders must comply with both the ethics opinion's framework and any applicable court-specific requirements.
Obligations Summary
| Rule | Obligation Area | AI-Specific Requirement |
|---|---|---|
| 4-1.1 | Competence | Understand AI tool capabilities and limitations; verify output before relying on it |
| 4-1.6 | Confidentiality | Conduct vendor due diligence before inputting client data; obtain informed consent where required |
| 4-5.1 / 4-5.3 | Supervision | Ensure AI-assisted work product is reviewed by a competent attorney; implement firm-level AI policies |
| 4-1.5 | Fees | Bill only for time actually spent; disclose AI tool costs if passed through to client |
| 4-1.4 | Communication | Disclose AI use to client where informed consent is required for confidentiality purposes |
Comparison with ABA Formal Opinion 512
The Florida opinion is broadly consistent with ABA Formal Opinion 512 (2024), which addressed generative AI under the Model Rules. Both opinions apply the same competence framework — understand the tool, verify the output — and both treat confidentiality as the primary risk requiring active due diligence rather than passive reliance on vendor assurances.
The Florida opinion goes somewhat further on billing, providing more explicit guidance on the time-compression issue. The ABA opinion addressed billing in general terms; Florida's version makes the reasonableness analysis more concrete for attorneys billing hourly.
Neither opinion requires disclosure of AI use to opposing counsel or courts as a standalone professional responsibility obligation — though both note that other rules or orders may create that requirement in specific contexts.
What This Opinion Does Not Cover
- Specific AI tools are not named or evaluated. The opinion sets a framework; attorneys must apply it to each tool they use.
- Unauthorized practice of law issues arising from AI-generated legal content are not addressed here (that is a separate UPL analysis).
- The opinion does not address AI use in non-legal administrative tasks (scheduling, billing software) — only tasks that constitute legal work.
- Court-specific AI disclosure requirements are outside the opinion's scope and must be checked separately for each jurisdiction and judge.
Comments
Join the discussion with an anonymous comment.