AI for Paralegals
The workflow that keeps your supervising attorney out of sanctions. 8 lessons grounded in ABA Opinion 512, Mata, Park, and the $110K Couvrette ruling. Certificate included.
On April 2026, in Couvrette v. Wisnovsky (D. Oregon), a federal judge imposed the largest AI hallucination sanction to date — $109,700 — against two attorneys whose briefs contained 15 fabricated case citations and 8 invented quotations. The local counsel, Tim Murphy of Portland, was sanctioned $14,000 specifically for “willful” failure to supervise. The lead attorney, Stephen Brigandi of San Diego, had attempted to conceal the errors when they were flagged.
The throughline across every AI sanctions case since Mata v. Avianca in 2023 is the same: a verification step was skipped. That step is almost always paralegal territory.
ABA Formal Opinion 512 (July 2024) made the framework explicit. Rule 5.3 imposes an affirmative supervision duty — attorneys who supervise nonlawyers cannot insulate themselves from a paralegal’s AI-generated errors. At the same time, paralegals who own the verification workflow become the firm’s first and most effective line of defense.
This course is the workflow. Eight lessons. First two free. Citation verification protocol, enterprise tool tour, Kovel-adjacent at-direction framework, and a one-page policy template your firm will actually adopt. Credential: PLG-XXXXXX.
What You'll Learn
- Explain ABA Formal Opinion 512 and its Rule 5.3 supervision implications for nonlawyers
- Identify the 5 paralegal workflows that create the highest privilege-waiver risk
- Apply a primary-source citation verification protocol that catches AI-hallucinated citations
- Implement a Kovel-adjacent AI workflow at supervising-attorney direction
- Evaluate enterprise legal AI tools (CoCounsel, Lexis+ AI, Bloomberg Law Brief Analyzer, Harvey, Spellbook)
- Draft a one-page firm AI use policy without sounding like a compliance scold
After This Course, You Can
What You'll Build
Course Syllabus
Prerequisites
- Completion of 'ChatGPT Without the Liability' recommended (not required)
Who Is This For?
- Paralegals at small and mid-size law firms
- Senior paralegals and legal operations managers
- Contract attorneys and document-review specialists
- Legal assistants transitioning into paralegal work
- Attorneys coaching their paralegal staff on AI practice
Frequently Asked Questions
Who is this course for?
US paralegals at small-to-midsize law firms. Senior paralegals, legal operations managers, and contract attorneys also benefit. Not for law students (different audience). Not for licensed attorneys (they're covered by ABA Opinion 512 directly and need firm-level training).
Is this legal advice?
No. It's a workflow course informed by ABA Formal Opinion 512, Mata v. Avianca, Park v. Kim, Couvrette v. Wisnovsky, and firm memos from Morgan Lewis, K&L Gates, and Jones Walker. For specific legal advice about your own matter or firm, consult a licensed attorney.
How is this different from 'ChatGPT Without the Liability'?
That course is the framework for any regulated professional. This course is the deep paralegal playbook. Citation verification, enterprise tool tour (CoCounsel, Lexis+ AI, Harvey, Spellbook), Rule 5.3 supervision specifics, firm policy drafting. Take Course 2 first if you haven't — it sets the foundation.
Will I get a certificate?
Yes — a verifiable PLG-XXXXXX certificate after all 8 lessons and the capstone. Shows you understand paralegal AI ethics and workflow at a specific, documented level most firms don't yet teach internally.
Is this still accurate if ABA issues a new opinion after 512?
Opinion 512's four-pillar framework (competence / confidentiality / supervision / candor) is the foundation; a future opinion would refine, not replace. The court cases (Mata, Park, Couvrette) remain primary evidence. Expect quarterly material updates as the case law evolves.
