ABA Opinion 512 vs Florida Bar 24-1: Reconciling the Two AI Ethics Frameworks

ABA Formal Opinion 512 and Florida Bar Opinion 24-1 both govern an attorney's use of generative AI, but they differ in emphasis. Here is how a Florida attorney reconciles competence, confidentiality, fees, and candor under both.

What ABA Formal Opinion 512 Actually Says

Issued by the American Bar Association’s Standing Committee on Ethics and Professional Responsibility on July 29, 2024, Formal Opinion 512 addresses the ethical obligations of lawyers when using generative artificial intelligence. The opinion does not establish new rules but instead interprets existing ABA Model Rules in the context of AI. Its primary focus is to ensure that the adoption of new technology does not compromise core professional duties.

The opinion maps existing rules to AI use, emphasizing a lawyer’s fundamental duties of competence (Rule 1.1), confidentiality (Rule 1.6), and proper supervision (Rules 5.1 and 5.3). It also touches on communication with clients (Rule 1.4), billing practices (Rule 1.5), and candor to the tribunal (Rule 3.3). The ABA's position is that while AI can be a valuable tool, the ultimate responsibility for the work product and professional conduct remains with the lawyer.

What Florida's Advisory Opinion 24-1 Requires

On January 19, 2024, The Florida Bar’s Board of Governors approved Ethics Advisory Opinion 24-1, which provides specific, Florida-centric guidance on generative AI. It largely aligns with the ABA's views but places a more explicit and pronounced emphasis on the duty of confidentiality under Florida Rule 4-1.6.

Florida's opinion directly confronts the risk of client data being used to train third-party AI models. It cautions that lawyers must obtain informed consent from clients before inputting any confidential information into a generative AI tool that lacks adequate confidentiality and security protections. This sets a high bar for using many popular, public-facing AI platforms. The opinion also reinforces duties of technological competence (Rule 4-1.1), supervision (Rule 4-5.3), and reasonable billing for AI-assisted work (Rule 4-1.5).

Rule 4-1.1: The Duty of Competence in the Age of AI

Both the ABA and Florida frameworks start with competence. ABA Model Rule 1.1 requires a lawyer to provide competent representation, which includes keeping abreast of changes in the law and its practice, "including the benefits and risks associated with relevant technology." Florida Rule 4-1.1 mirrors this language.

In practice, this means you cannot simply use an AI tool without understanding how it works. This isn't about knowing how to code a large language model. It's about understanding its limitations.

• Hallucinations: Generative AI can invent facts, cases, and citations. A competent attorney must independently verify every factual assertion and legal citation produced by an AI. The sanctions issued in the Southern District of New York case *Mata v. Avianca, Inc.* serve as a standing warning against submitting AI-generated filings without rigorous verification.
• Bias: AI models are trained on vast datasets from the internet, which contain inherent biases. An attorney must review AI-generated drafts for biased language or skewed analysis, particularly in sensitive practice areas like employment or family law.
• Scope: The AI's knowledge is limited to the data it was trained on. A 15-attorney real estate firm in Orlando using an AI to analyze a new local ordinance passed in Q2 2026 must recognize that a general-purpose model trained on data from 2024 will have no knowledge of it. Competence requires supplementing AI output with current, jurisdiction-specific research.

The duty of competence under both opinions requires a hands-on, skeptical approach. The lawyer is the final arbiter of the work product's accuracy and validity, not the AI.

Rule 4-1.6: Confidentiality is the Sticking Point

Here lies the most significant operational difference between the two opinions, with Florida's 24-1 being more explicit. ABA Rule 1.6 and Florida Rule 4-1.6 both demand that a lawyer "shall not reveal information relating to the representation of a client." The key question is whether inputting case information into a cloud-based AI constitutes "revealing" it.

For most commercial AI tools, the answer is yes. Their terms of service often grant the provider a license to use submitted data to train their models. This is a non-starter for client confidentiality.

Consider a 12-attorney personal injury firm in Tampa. Their case files contain extensive protected health information (PHI) subject to HIPAA, detailed accident reports, and privileged client communications. Uploading a client's medical records or deposition summary to a public cloud AI service to "summarize the key injuries" would almost certainly violate Rule 4-1.6. Florida Opinion 24-1 suggests this requires explicit, informed client consent, which itself presents a complex ethical hurdle.

The ABA opinion is slightly less prescriptive, advising lawyers to "make reasonable efforts to prevent the inadvertent or unauthorized disclosure of, or unauthorized access to, information relating to the representation of a client." Using a cloud platform without understanding its data policy would not meet this "reasonable efforts" standard. The core issue is the same: client data leaving the firm's direct control.

Rules 4-5.1 & 4-5.3: Supervising Your New AI "Assistant"

Just as a managing partner must supervise a junior associate or paralegal, attorneys must supervise their AI tools. ABA Rules 5.1 and 5.3, and their Florida counterpart 4-5.3, require partners and supervising lawyers to make reasonable efforts to ensure the firm has measures in place giving reasonable assurance that the conduct of nonlawyers is compatible with the professional obligations of the lawyer.

This means you cannot delegate a task to an AI and accept the output without review.

• Workflow Integration: The firm must have a documented policy for how AI is used. For example, an intake paralegal at an immigration practice might use an AI to perform an initial summary of a client's travel history. A supervising attorney must review that summary against the source documents before it is used in any filing.
• Training: Attorneys and staff must be trained on the firm's AI policy, the specific tools in use, and their limitations. This is a direct extension of the duty of competence under Rule 1.1.
• Accountability: The supervising lawyer is responsible for any errors or ethical breaches resulting from the use of AI by a subordinate, just as they would be for a human assistant. The AI is a tool, not a separate entity to which blame can be assigned.

Compliance Checklist: ABA 512 & Florida 24-1

This table provides a scannable summary for managing partners and IT directors to assess their AI posture.

Ethical Duty	ABA Opinion 512 Requirement	Florida Opinion 24-1 Requirement	Compliant Action Item
Competence (Rule 1.1/4-1.1)	Understand AI's limitations, including hallucinations and bias. Verify all outputs.	Stay current on technology risks and benefits. Ensure AI use is accurate.	Mandate verification of every AI-generated legal citation and factual claim.
Confidentiality (Rule 1.6/4-1.6)	Make "reasonable efforts" to prevent unauthorized disclosure of client data to third-party AI vendors.	Obtain client's informed consent before inputting confidential data into most third-party AI tools.	Use only on-premise, local AI tools that do not transmit client data to the cloud.
Supervision (Rule 5.1/5.3)	Supervise nonlawyers (and AI tools) to ensure their use conforms to a lawyer's professional obligations.	Implement policies and provide training for AI use; the supervising lawyer remains responsible for all work product.	Draft and enforce a firm-wide AI acceptable use policy. Appoint a partner to oversee compliance.
Billing (Rule 1.5/4-1.5)	Do not charge clients for the time it takes to learn to use an AI. Bill for actual legal work, not AI overhead.	Fees must be reasonable. Do not charge hourly rates for tasks completed in minutes by an AI.	Bill based on value delivered and actual attorney time spent reviewing/editing AI output. Consider flat fees for AI-assisted tasks.
Candor (Rule 3.3/4-3.3)	A lawyer is responsible for the accuracy of all statements made to a court, regardless of their origin.	Do not submit AI-generated content containing false statements or citations to a tribunal.	Implement a "human-in-the-loop" protocol where an attorney must approve every sentence in a court filing.

What Mi Assist Legal Does

Our team at Mi Assist Legal built OpenClaw Legal specifically to address the confidentiality problem at the heart of both ABA Opinion 512 and Florida Opinion 24-1. Our system installs a dedicated AI server, typically a Mac Mini, inside your firm's existing network. It does not send your case files, your queries, or any other client data to the cloud. Ever.

OpenClaw Legal uses open-source models like Llama 3.2 running on local hardware via Ollama, and it indexes your firm's documents into a private ChromaDB or Qdrant vector store on that same local server. This allows your attorneys to ask natural-language questions of your entire case history while maintaining full compliance with your duties under Rule 1.6 and 4-1.6.

Frequently Asked Questions

Is an on-premise AI system difficult for a small firm to manage?

No. Our OpenClaw Legal appliance is designed as a "set it and forget it" solution. Our team handles the initial installation, configuration, and document indexing. Ongoing maintenance is minimal and can be managed by your existing IT person or MSP, with our team providing direct support.

Does using a local AI like OpenClaw satisfy my duty of competence?

It helps, but the duty remains with you. OpenClaw provides source-cited answers, linking every statement back to the specific documents in your case files it used for the response. This makes verification, a key part of competence, significantly faster and more reliable than with a public AI. However, you are still the attorney of record and must exercise professional judgment.

How does billing for work done with OpenClaw work?

Under ABA and Florida rules, you cannot bill an hour for a task an AI completes in two minutes. However, you can and should bill for the attorney time spent crafting the query, reviewing the AI's output, verifying its sources, and integrating it into the final work product. OpenClaw makes your team more efficient, allowing you to produce higher-quality work faster, which supports value-based or flat-fee billing models and increases firm throughput.

Can we choose which documents the AI has access to?

Yes. During setup, we work with you to define the scope of the document index. You can limit it to specific practice groups, date ranges, or matter types. Access controls can be configured to ensure that attorneys can only query documents they are authorized to view, mirroring your existing document management permissions.

Schedule a 30-minute on-site assessment with our team. We will evaluate your firm's network and document management system and provide a clear plan for deploying a secure, ABA Rule 1.6-compliant AI assistant.