Module 3.1 · Topic 5

AI Literacy for Legal Professionals

Bottom Line Up Front: Legal professionals have a professional obligation to understand generative AI, its capabilities and limitations, before using it in client work. This is not optional technical knowledge; there is…

5.1 Why Legal Professionals Need Technical AI Understanding

ABA Formal Opinion 512 establishes that lawyers using AI must understand the technology.

Without technical understanding, you cannot supervise effectively. You won't know which tasks are safe to delegate or which need verification. You won't recognize hallucinations, know when knowledge cutoffs matter or know what model behaviors correspond with different degrees of reliance. Informed use means understanding limitations well enough to decide: do I need to verify this output? If so, how thoroughly? Use that understanding to calibrate verification and protect clients.

5.2 Key Concepts That Impact Legal Practice

You've learned several major concepts about how generative AI works. But not all concepts matter equally for all practice areas. A litigation attorney's priorities differ from a transactional attorney's. A regulatory specialist's risks differ from an intellectual property lawyer's. This prompt helps you identify which concepts to prioritize for your specific practice.

Logic behind this approach:

Mastering every AI detail is overwhelming. Identify which concepts matter most for your practice. Researchers prioritize citation hallucination risk; regulatory specialists prioritize knowledge cutoff limitations; litigators prioritize tool-use for evidence research. Focused learning beats scattered learning.

Sample prompt:

I practice [AREA]. My AI use cases include [MAIN USES]. Which concepts are most critical? 1. Hallucination in citations 2. Knowledge cutoff limits 3. Context window limits 4. Pattern-matching vs. knowledge tasks 5. Tool-use and external data 6. Recognizing when AI guesses For each, explain risks in my practice and verification strategies.

What to expect:

A good response ranks concepts by relevance, explains which risks matter most, and suggests verification strategies. Use to focus your learning and team governance conversations. Unsatisfied with the answer? Enhance the context with an example of your daily work that you'd like to safely support with AI.

5.3 The Hallucination Problem in Legal Context

Mata v. Avianca, and many cases since, established that lawyers relying on AI-generated citations without verification will likely face sanctions. Fabricated statutes or holdings breach competence duties and create malpractice liability.

Critical Risk: Hallucinations in Legal Work

The specific danger: Generative AI can produce plausible-sounding but entirely fabricated case citations, statutory language, regulatory interpretations, and legal analysis. Because these outputs sound authoritative and are presented with confidence, they can easily slip into client work, court filings, or legal opinions without detection.

Process for assessing hallucination risk:

Classify: Pattern-matching (lower-risk) or knowledge task (higher-risk)?
Identify stakes: If AI gets this wrong, what's the consequence?
Check external knowledge: Does it require facts beyond provided documents? Find them.
Verify and document: For high-stakes, verify every factual claim. Record your process.

5.4 Professional Responsibility and AI Competence

Your professional rules require you to provide competent representation and keep client information confidential. Using AI in ways that violate these rules without understanding the tool's limitations, or without verifying output, or by uploading sensitive client information to an unvetted system creates liability. This prompt helps you assess whether you're meeting your competence obligation.

Logic behind this approach:

Model Rule 1.1 (Competence) requires understanding technologies you use. Rule 1.6 (Confidentiality) requires protecting client information. Before using AI in client work, assess whether you meet these obligations.

Sample prompt:

I'm using [AI TOOL] for [TASK]. Assess my professional obligations: 1. Competence: Do I understand this AI's training data, cutoffs, hallucination risks, limitations? 2. Confidentiality: Where is data stored? Could it train the model? Security adequate? 3. Supervision: How will I verify output? What should I document? 4. Client communication: Must I disclose AI use? What limitations should clients know? What steps must I take before using this in client work?

What to expect:

A thorough response identifies competence gaps, flags confidentiality risks, recommends verification strategies, and suggests client disclosures. Document your analysis before deploying any AI in client work.