Sources Cited
Public summaries of cases, policy materials, and AI-governance sources behind the CounselCore position on sovereign in-house legal AI.
These materials support the core CounselCore argument: law firms and legal departments need AI systems that preserve privilege, keep evidence in their custody, and avoid exposing sensitive matter data to uncontrolled cloud workflows.
The Tenth Circuit proposes human verification of AI-assisted filings, covering authorities, quotations, citations, legal analysis, and rule compliance. The proposal is not yet in force as of October 15, 2026.
New York judicial ethics guidance permits approved private AI to assist with transcript summaries, subject to required training, reliability and neutrality review, and the judge’s continuing responsibility for decisions.
NYC Bar guidance addresses AI recording beyond client calls, including participant permission, storage and sharing, and tactical risks. It recommends avoiding recording by default absent a good reason.
The New Jersey Judiciary maintains a centralized collection of AI principles, guidance, lawyer notices, educational materials, and responsible-use resources.
The ABA discusses sanctions arising from fabricated AI-generated authorities and reinforces the lawyer's duty to verify every citation, quotation, and proposition submitted to a court.
ILTA treats AI readiness as an information-governance decision: document cleanup, metadata, taxonomy, security, permissions, source quality, and ongoing corpus maintenance determine whether retrieval can be trusted.
The article treats unauthorized AI as an enterprise risk involving sensitive data, unreviewed providers, policy violations, and records outside normal security and retention controls.
The integration combines external legal content with permission-aware firm knowledge, raising the competitive standard for source provenance, ethical walls, and knowledge reuse.
A practical summary of Anthropic’s Fable 5 and Mythos 5 access statement and what it signals about AI governance, model access, and institutional risk.
The most striking example is LNU v. Blanche, Ninth Circuit, June 3, 2026. Attorneys Mike Singh Sethi and William Rounds submitted briefs containing: nonexistent cases; misattributed quotations; grossly misrepresented real cases. They repeatedly denied…
The article teaches lawyers to treat an unusually perfect AI-generated case as a verification trigger and explains that grounding improves traceability without guaranteeing legal accuracy.
This is one lawyer who generated four separate state disciplinary consequences from one AI episode. The underlying conduct Matthew Brett Reeves used ChatGPT to add five citations to two motions in federal court in…
A practical summary of Miller v. Regions Bank and its lessons for AI-generated legal research, false authority, candor, and evidence preservation.
Indiana provides an interesting contrast. Attorney Tae Sture filed a brief containing two nonexistent authorities generated through an AI-enabled legal research workflow involving Fastcase. The magistrate initially recommended $7,500. The district judge ultimately reduced…
The commentary frames cybersecurity as a leadership, professional-responsibility, vendor-risk, monitoring, incident-preparation, and client-service concern rather than an isolated IT function.
This is another particularly significant case. Janelle Melissa Lewis — New York Lewis was hired through an online freelance-lawyer platform by a Texas immigration attorney to prepare a response to an order to show…
The article explains how ordinary prompting can expose client facts, documents, or strategy to externally hosted systems and recommends approved platforms, policy, training, and provider-term review.
The article describes an AI-native Newark office intended to advance firmwide AI tools, internal governance, and client advisory work. It shows legal AI moving from isolated experimentation into law-firm operating strategy.
This is another unusually severe case. Deborah Leslie — Georgia Leslie was an assistant district attorney in Clayton County prosecuting the Hannah Payne murder case. She used AI in preparing legal filings. The filings…
The research benchmarks legal retrieval-augmented generation and finds meaningful improvement from specialized methods while retrieval and reasoning failures remain material.
ACC's second-edition toolkit organizes AI adoption around maturity, governance, use cases, ethics, intellectual property, vendor contracting, outside counsel, and emerging agentic systems.
The collaboration illustrates market demand for permission-aware AI that retrieves and reasons over law-firm institutional knowledge rather than relying only on general information.
The article connects AI adoption to junior-lawyer training, validation skills, staffing, leverage, pricing, and the need to preserve distinctly human legal judgment.
The ABA review reports widespread personal AI use alongside concerns about security, ethics, privilege, and trust, while many organizations still lack training and enforceable policy.
This is currently the case I would regard as the most important precedent in the United States. W. Gregory "Greg" Lake — Nebraska What happened Lake represented a father in Prososki v. Regan, a…
Survey data from about 1,300 legal-industry respondents shows broad use of general and legal-specific AI, but far fewer organizations with enforced policy or comprehensive training.
The analysis describes a communication gap between law firms and legal departments over AI use, client expectations, billing, and return on investment.
Arizona has also now produced a genuine state disciplinary case involving AI. John A. Griffiths — Arizona The state disciplinary record identifies use of Eve.legal, with fabricated/misrepresented case authority. The resulting discipline was a…
In Whiting v. City of Athens, the Sixth Circuit imposed another unusually severe sanction. The attorneys' briefs contained more than two dozen fake citations and factual misrepresentations, with hallmarks of AI hallucination. The court…
The analysis describes tension between clients seeking AI-enabled efficiency and transparency and firms trying to price, measure, and capture the value of technology investment.
The ABA compares federal decisions showing that AI communications are not automatically privileged, automatically waived, or automatically work product; purpose, confidentiality, counsel direction, and disclosure matter.
The article explains that AI-generated meeting transcripts, summaries, chats, and metadata can become electronically stored information subject to retention, litigation holds, and discovery.
A practical, SEO-friendly summary of United States v. Heppner and what it means for generative AI, attorney-client privilege, work product, and law firm AI governance.
Omid Zareh — New York This one is much cleaner as an AI-related disciplinary precedent. Zareh participated in preparation/review of a federal brief containing: numerous citation errors; misrepresented case law; authorities that did not…
Thomson Reuters documents a sharp rise in organization-wide AI use while systematic return-on-investment measurement remains limited. Adoption is becoming normal faster than many organizations can evaluate it.
The ABA checklist translates ethical duties into an operational workflow: use approved systems, protect information, verify sources and reasoning, train users, and require final human signoff.
The ACC resource gives in-house teams a due-diligence structure for evaluating outside counsel's approved tools, data handling, human review, governance, records, incident readiness, and value.
ACC reports that corporate legal-department AI use more than doubled, while many clients still did not know whether outside firms were using AI or adjusting pricing for AI-enabled efficiency.
New Jersey experts emphasize risk assessment, pilot programs, employee upskilling, transparency, security, and flexible governance as organizations implement AI.
ACC's public resource describes sample outside-counsel AI guidelines addressing disclosure, data security, accuracy, and performance expectations.
Steven J. Marullo — Massachusetts This was a particularly instructive case. Marullo represented an estate in a wrongful-death case. His associate and interns prepared opposition papers. The associate used AI for legal research. The…
A practical summary of Kohls v. Ellison and its lessons for expert declarations, deepfake litigation, AI hallucinations, and evidentiary reliability.
The article frames AI agents as emerging legal teammates that depend on structured firm knowledge, matter context, workflow boundaries, governance, and human oversight.
The article emphasizes AI and vendor inventories, data minimization, contractual safeguards, and firmwide governance, including AI features introduced through existing software.
ABA Formal Opinion 512 applies existing duties of competence, confidentiality, client communication, supervision, candor, and reasonable fees to generative AI use.
The article describes a cross-disciplinary AI practice group addressing regulation, intellectual property, privacy, cybersecurity, transactions, and ethics.
A practical summary of Kruse v. Karlen and its warning about fictitious AI-generated cases, appellate briefing, and frivolous appeals.
A practical summary of Park v. Kim and what it teaches about AI-generated citations, appellate filings, and professional responsibility.
Opinion 24-1 permits generative AI use while emphasizing confidentiality, output verification, supervision, reasonable fees, and compliance with advertising rules.
The early article documented concerns about confidentiality, accuracy, liability, limited formal policy, and the need for human oversight and training before current adoption levels became widespread.
A clear summary of Mata v. Avianca, the landmark sanctions order involving ChatGPT-generated fake cases and the legal duty to verify AI research.
Use these source summaries as a starting point for a confidential briefing on in-house AI, privilege, evidence custody, and governance controls.
