ABA (National) — Formal Opinion 512
Generative Artificial Intelligence Tools
The ABA's first formal ethics guidance on generative AI in law practice. Clarifies that existing ethical obligations — competence, confidentiality, communication, candor, supervision, and fees — fully apply when attorneys use AI tools.
Key Takeaways
- Lawyers must independently verify all AI-generated outputs including legal citations
- Client confidential information must not be entered into AI tools without adequate security review
- Informed consent may be required before using AI on client matters
- Billing must reflect actual time spent, not inflated hours from pre-AI workflows
- Firms should establish clear AI usage policies and train all staff
★ Landmark Opinion
Kentucky — KBA E-457
AI Use in Legal Practice
States that lawyers must stay informed about AI developments in legal practice. Routine AI use for legal research generally does not require disclosure, but fees should be reduced if AI leads to significant efficiency gains.
Key Takeaways
- Routine AI legal research does not require client disclosure
- Reduce fees when AI significantly reduces actual work time
- Lawyers must stay current on AI capabilities and risks
North Carolina — Formal Ethics Opinion 1
AI Competence and Billing
Requires lawyers to use AI competently and securely, exercise independent judgment, and properly supervise its use. Lawyers may not inaccurately bill clients based on 'time-value represented' if AI reduced the actual work time.
Key Takeaways
- Independent judgment must always be exercised over AI outputs
- Cannot bill pre-AI time rates when AI has reduced actual effort
- Competent use requires understanding AI tool limitations
New Mexico — Ethics Advisory Opinion 2024-004
AI Tools and Ethical Obligations
Highlights the need for lawyers to understand AI tool parameters, avoid inputting confidential information into insecure tools, and verify all AI outputs for accuracy.
Key Takeaways
- Understand parameters and limitations of every AI tool used
- Never input confidential data into insecure AI platforms
- All AI outputs must be verified for accuracy before use
Oregon — Formal Opinion No. 2025-205
AI in Legal Practice
Comprehensive survey of AI ethics issues covering competence, confidentiality, billing, supervision, and appropriate use in court filings. Closely aligns with ABA Formal Opinion 512.
Key Takeaways
- All six ABA ethical categories apply to AI use in Oregon
- Court filings generated with AI assistance require thorough verification
- Supervisory obligations extend to AI tool use by non-lawyer staff
Texas — Opinion 705
Application of DRPC to AI Usage
Applies existing Texas Disciplinary Rules of Professional Conduct to AI usage, providing a framework for competence, confidentiality, verification of AI-generated information, and fair billing practices.
Key Takeaways
- Existing disciplinary rules fully apply to AI — no new rules needed
- Verification of AI outputs is non-negotiable before reliance
- Fair billing requires transparency about AI-assisted efficiency
California — Practical Guide
Attorney Competence with LLMs
Published practical guidance emphasizing that attorney competence requires understanding large language models, including their risks like hallucinations and data privacy concerns.
Key Takeaways
- Attorneys must understand how LLMs work, including hallucination risks
- Data privacy must be maintained when using cloud-based AI tools
- Competence is an ongoing obligation — requires keeping current with AI evolution
Pennsylvania — Joint Formal Opinion 2024-200
AI-Generated Legal Work
Warns that AI can generate work product but does not replace the professional duty to verify all case law references and legal analysis. Attorneys remain fully responsible for all submitted work.
Key Takeaways
- AI does not replace the duty to verify all case law citations
- Attorneys bear full responsibility for AI-assisted work product
- Review and verification are non-delegable professional duties
SDNY (Federal Court) — Mata v. Avianca, Inc.
Sanctions for AI-Fabricated Citations
Landmark case where an attorney was sanctioned for submitting a brief containing fictitious case citations generated by ChatGPT. The court imposed sanctions for violating the duty of candor and competence.
Key Takeaways
- Attorney sanctioned for submitting AI-fabricated case citations to federal court
- Court held that AI-generated work must be verified just like any other legal research
- Became the leading cautionary example for AI use in litigation
★ Landmark Opinion
