Wilder v. Harvard: Massachusetts Denies Pro Hac Vice Admission Based on Attorney’s Prior AI Sanction — MX2.law Named

Background: A Harvard Lawsuit and a Florida Attorney’s Problematic Past

Glenn Wilder and other plaintiffs were suing Harvard College in consolidated actions managed by appointed lead counsel. Attorney T. Michael Morgan of Morgan & Morgan’s Orlando office sought permission to appear in the Massachusetts proceeding pro hac vice. Morgan was not a member of the Massachusetts bar. Three other Morgan & Morgan attorneys were already appearing for the plaintiffs.

In disclosing prior disciplinary history, Morgan revealed he had been sanctioned in Wyoming federal court for signing motions containing cases hallucinated by Morgan & Morgan’s in-house AI platform. This disclosure — required by bar rules — became the basis for denying his application.

The AI Issue: MX2.law — A Named In-House AI Platform as Sanctioned Tool

The Wyoming case (Wadsworth v. Walmart Inc., 348 F.R.D. 489 (D. Wyo. Feb. 24, 2025)) identified Morgan & Morgan’s in-house AI platform — which the firm calls “MX2.law” — as the source of eight non-existent case citations in motions in limine. Morgan’s associate drafted the motions using MX2.law; Morgan and local counsel affixed their e-signatures without reviewing the documents. Judge Rankin found that Morgan had violated Rule 11(b) by signing filings he had not read and that cited cases “hallucinated by an AI platform.”

The Massachusetts Superior Court found that this prior AI sanction constituted a “recent ethical lapse” sufficient to deny pro hac vice admission. The court was also concerned about Morgan’s decision to file in Massachusetts without permission (violating G.L. c. 221 § 46A and SJC Rule 3:15). Combined, these two failures — prior AI sanction and immediate procedural violation in the current case — were sufficient to deny the application.

Holdings

1. Pro hac vice application denied. Attorney Morgan’s prior AI sanction, combined with his procedural violation in filing without permission in Massachusetts, provided ample grounds for denial.
2. AI sanctions travel with the attorney. This is the clearest statement in the published case law that AI hallucination sanctions have bar admission consequences beyond the immediate proceeding. An attorney sanctioned for AI citations in one court can be denied admission to another court on that basis alone.
3. Named AI platform documented. “MX2.law” — Morgan & Morgan’s proprietary AI platform — is now identified by name in a published judicial order as the source of hallucinated case citations. This is the first time a law firm’s internal AI tool has been named in a published judicial proceeding.
4. Nondelegable review duty confirmed. Even reliance on an in-house AI platform developed by the firm does not excuse the reviewing attorney’s failure to read filings before signing them.

“Attorney Morgan’s role in submitting hallucinated case citations to a Federal District Court, and his failure in that action to review motions that he allowed to be filed over his signature, provide ample grounds for denying his request to appear pro hac vice in this case.”

— Massachusetts Superior Court, Business Litigation Session, Wilder v. President & Fellows of Harvard College

India Angle: Prior AI Sanctions and Bar Council Fitness Assessments

The Wilder/Harvard case introduces the most consequential dimension of AI hallucination liability: the fact that prior AI sanctions can be used to deny future court access — effectively a professional reputation consequence that follows the advocate. In India, the equivalent mechanism is the BCI Enrollment Certificate and “good standing” certification system under the Advocates Act.

Relevant Indian Law

• Advocates Act 1961 Section 30 / BCI Rules: An advocate appearing in a court outside their enrolled state must obtain permission. Courts and tribunals in other states have discretion in granting such permission. A prior BCI disciplinary order — the Indian equivalent of an AI sanction — could be used to deny cross-bar appearance permission, just as Wyoming’s AI sanction was used to deny pro hac vice in Massachusetts.
• Bar Council of India Rule 33 / Disciplinary Records: BCI disciplinary records are accessible to courts and other Bar Councils. A State Bar Council’s disciplinary order against an advocate for AI-related misconduct would follow that advocate when they seek to appear before courts in other states or before the Supreme Court.
• In-house AI tools in Indian law firms: Several Indian law firms are developing or procuring AI-assisted drafting tools. The Morgan & Morgan “MX2.law” precedent should be a warning: an in-house AI tool that generates hallucinated citations, when signed off on by an attorney who has not verified them, creates Rule 11 liability. Indian firms must implement mandatory verification protocols for any AI-generated brief drafts — and ensure that the signing advocate personally verifies every citation.

Three Practical Tips for Indian Practitioners

1. AI sanctions are not local — they follow you. An AI-related sanction, BCI disciplinary order, or judicial finding of AI misconduct will be disclosed in any good standing certificate or fitness-for-practice assessment. Treat every AI citation error with the same seriousness as any other professional conduct matter, because it will appear in your record.
2. When your firm uses an in-house AI tool, the signing advocate bears personal responsibility. The “it was our firm’s AI platform, not me personally” defence was expressly rejected by the Wyoming court and confirmed by Massachusetts. If you sign a brief, you are certifying that you have read it and its citations are accurate — regardless of how it was drafted.
3. E-signatures on AI-generated documents without reading them are per se Rule 11 violations. Morgan affixed his e-signature to motions he never read. In India, vakalatnama-based filings bear the advocate’s signature on the court record. If you sign a document without reading it — even if drafted by AI or a junior — you are personally responsible under BCI Rules and the Advocates Act for every statement in that document.

Quick Takeaways

• Prior AI sanctions can be used to deny future pro hac vice admissions — establishing that AI hallucination consequences travel across courts and jurisdictions.
• A named in-house law firm AI platform (“MX2.law”) has been identified by a court as the source of hallucinated case citations — the first such firm-level AI attribution in published case law.
• Signing a motion without reading it is an independent Rule 11 violation, separate from the AI hallucinations themselves — the duty to read what you sign is absolute and nondelegable.
• Morgan & Morgan’s pattern — a large firm’s associate using AI, a partner signing without review — is likely common in large law firms and represents a systemic risk beyond individual liability.
• In India: BCI disciplinary records follow advocates across bar councils; AI-related misconduct findings could affect cross-state appearance permissions and Supreme Court admissions.

Deep Dive: The Law Firm AI Platform Liability Problem

The Wilder/Wadsworth AI saga is among the most significant in the developing case law because it identifies a systemic law firm problem: a large personal injury firm (Morgan & Morgan, with offices in multiple states) developed or procured an in-house AI platform that generated hallucinated case citations. Associates used this platform without adequate supervision. Partners signed briefs without reading them. The result was an AI sanction that followed the firm’s lead partner into a subsequent pro hac vice application in a different state court — creating cascading consequences.

For Indian law firms developing or deploying AI drafting tools, the Wadsworth/Wilder sequence is a warning about institutional AI governance: (1) AI tools must have verifiable citation accuracy before deployment; (2) associates who use AI tools must personally verify all citations before submitting drafts to partners; (3) partners who sign filings must personally verify citations — they cannot delegate this duty upward to AI or downward to associates; and (4) AI sanctions in one court create discoverable records that may affect future practice opportunities across courts and jurisdictions.