ABA Ethics &
AI Compliance
How Legience addresses ethical obligations under ABA Formal Opinion 512, applicable Rules of Professional Conduct, and emerging AI ethics standards for legal practice.
ABA Formal Opinion 512: Generative AI Tools
In July 2024, the American Bar Association issued Formal Opinion 512, providing guidance on attorneys' ethical obligations when using generative AI tools. Legience is designed to help attorneys meet their obligations under this opinion.
1. Competence (Rule 1.1)
Attorneys must understand the AI tools they use sufficiently to assess their reliability and limitations.
- Case-law citations are automatically checked against the CourtListener database, and statute and rule citations link to official government sources; citations that can't be verified are flagged rather than linked.
- AI outputs are clearly labeled as AI-generated, and every response reminds you to verify all citations against primary sources before relying on them.
- We provide documentation on how our AI works, its limitations, and best practices for use.
- LegiSearch™ flags the legal treatment of each cited source (good law, distinguished, or overruled) and shows how many citations are independently verified — prompting you to confirm treatment against Shepard's or KeyCite before filing.
2. Confidentiality (Rule 1.6)
Attorneys must ensure that client information shared with AI tools remains confidential.
- Secure AI Processing: AI features powered through AWS Bedrock under our Business Associate Agreement (BAA).
- No AI Training: Attorney-client data is never used to train, fine-tune, or improve any AI model.
- Encryption: Encrypted in transit (TLS) and at rest via AWS KMS storage-layer encryption, with additional AES-256 field-level encryption for selected sensitive data.
- US-Only Processing: All data remains in AWS US-East. No offshore processing.
- Conversation Management: AI conversation history is stored in our database for continuity. You can delete individual conversations at any time.
3. Supervision (Rules 5.1 and 5.3)
Attorneys must supervise AI outputs as they would supervise a junior associate or paralegal.
- AI-generated documents are saved as drafts, never as final versions, and carry a persistent DRAFT watermark until an attorney reviews and approves them.
- LegiDraft™ documents go through an attorney review-and-approval workflow (approve & finalize, or request changes) before they're finalized; editing an approved document reverts it to draft for re-review.
- When an attorney reviews, approves, or requests changes to an AI-generated document, the audit log records who acted and when, as a tamper-evident entry.
4. Communication (Rule 1.4)
Attorneys should communicate with clients about the use of AI in their matters when appropriate.
- Legience includes an optional, standalone AI-use disclosure that your firm can send clients to review and acknowledge; the acknowledgment is recorded on the client's record (ABA Model Rule 1.4).
- An AI-disclosure notice is also included in the public intake form.
5. Fees (Rule 1.5)
Attorneys must ensure that fees for AI-assisted work are reasonable and transparent.
- AI-generated documents record both a generation time and a separate attorney review time, so you can account for the work accurately.
- Time tracking supports billable and non-billable hours by attorney, case, and rate, with a review-and-approval step before invoicing.
State-Specific Compliance
In addition to ABA guidance, Legience addresses jurisdiction-specific requirements:
- 201 CMR 17.00: Security practices designed to align with the Massachusetts Standards for the Protection of Personal Information — including access controls, encryption, and a documented incident-response and breach-notification process. (A formal Written Information Security Program and independent certification are in progress.)
- applicable Rules of Professional Conduct: Our platform is designed to support compliance with Rules 1.1, 1.6, 1.15, 5.1, and 5.3 as they apply to AI-assisted legal practice.
- Conflict Checking: Automated conflict-of-interest screening across clients, matters, and adverse parties — multi-type detection, review and waiver workflows, and waiver-document tracking (informed by ABA Model Rule 1.9). Results are advisory and support, but do not replace, your professional judgment.
Our Commitment
Legience is built by people who understand that attorney-client privilege is sacred and that ethical AI use requires more than marketing promises — it requires architectural decisions. Our secure AI processing through AWS Bedrock, encryption standards, and audit capabilities are designed to make ethical AI compliance the default, not an afterthought.
Questions?
For questions about our ethics compliance: ethics@legience.com or legal@legience.com.