Corrections & fabrications watch

How this page works

Every item below meets three evidence requirements before it is listed. First, we re-fetched the cited page live on the access date shown and quote only wording that was on the page that day. Second, each item links an independently archived copy of the page in the Internet Archive’s Wayback Machine, so readers can verify the quoted wording without taking our word for it. Third, where the issue is whether an opinion exists, we checked the issuing body’s own published opinion index and link it with the access date. We document pages, not people: an entry here means a specific published claim did not match the primary record on the date checked, and nothing more. We do not speculate about how any error was produced. Competitor sites are named only to identify where a cited page was published. If a page is corrected at the source, we verify the fix and update this log.

1. “Utah Bar Advisory Opinion 24-03”: an apparently fabricated citation

The AI Vortex page at aivortex.io/legal/ethics/utah/ (page dated April 11, 2026) states that “The Utah State Bar issued Advisory Opinion 24-03 in May 2024, providing targeted guidance on generative AI use in legal practice,” labels it a “Formal Opinion,” and builds billing, confidentiality, and consent guidance on it — but the citation appears to be fabricated.

No such opinion appears in the Utah State Bar’s published ethics-opinion index as of July 1, 2026 (utahbar.org/ethics-opinions/); the citation appears to be fabricated. The Bar’s published opinions include no 2024 advisory opinion of any number and no numbered opinion addressing generative AI; the most recent numbered opinion we could locate anywhere on utahbar.org is Ethics Opinion 23-01, issued August 10, 2023, on an estate-planning question (opinion PDF on the Bar’s site; archived copy of the Bar’s opinion-index page as checked — the index page itself lists nothing newer than 2022, while newer opinion PDFs appear elsewhere on the Bar’s site). The Utah State Bar’s published generative-AI materials as of our latest review are educational articles and standing-committee content, not a numbered formal ethics opinion.

Source page as published July 1, 2026 (archived copy: Wayback Machine). Absence check: Utah State Bar ethics-opinion index, accessed July 1, 2026.

2. “ISBA 24-01” as a generative-AI opinion: an apparently fabricated citation

The AI Vortex page at aivortex.io/legal/ethics/illinois/ (page dated April 11, 2026) states that “The Illinois State Bar Association issued ISBA Advisory Opinion 24-01 in August 2024, making Illinois one of the states with formal AI ethics guidance for attorneys,” and describes it as a “Formal Opinion” covering competence, confidentiality, billing, and disclosure — a citation that, as an August 2024 generative-AI opinion, appears to be fabricated.

The ISBA’s published opinion index does list a real Opinion 24-01, but that opinion was issued in March 2024 and addresses an Illinois lawyer acting as local counsel in a criminal matter for an out-of-state lawyer, indexed under court obligations and scope of representation rather than anything AI-related — and because no ISBA advisory opinion addressing generative AI appears anywhere in the index as of July 1, 2026, the August 2024 generative-AI opinion the page describes appears to be fabricated (ISBA Ethics Opinions by Year). Illinois does have real, verifiable instruments — the Illinois Supreme Court’s policy on artificial intelligence and the ARDC’s implementation guide, both listed in the tracker with primary sources — but as of our latest verification Illinois has no state bar formal ethics opinion on generative AI.

Source page as published July 1, 2026 (archived copy: Wayback Machine). Absence check: Illinois State Bar Association published opinion index, accessed July 1, 2026.

3. A West Virginia “no guidance” claim the primary record contradicts

The AI Vortex page at aivortex.io/legal/ethics/west-virginia/ (page dated April 11, 2026) states: “The West Virginia State Bar has no formal opinion, task force, or advisory report addressing generative AI in legal practice.”

The primary record contradicts this. The West Virginia Lawyer Disciplinary Board issued Legal Ethics Opinion 24-01, “Artificial Intelligence,” in June 2024. It is published by the Office of Lawyer Disciplinary Counsel, whose opinion archive lists it as “L.E.O. 2024-01 Artificial Intelligence” (wvodc.org/Legal-Ethics-Opinion), and it is in our tracker as a formal opinion. It is also one of the strictest instruments in the country on client consent: it requires informed client consent, confirmed in writing, before generative-AI use. Two precision notes. First, the issuer is the Lawyer Disciplinary Board, not the West Virginia State Bar itself, so a careful sentence about West Virginia names the Board. Second, the primary source publishes a month, June 2024, so we do not cite a day.

The page also does not agree with itself. After stating that no opinion exists, its own FAQ answers the question “What does the West Virginia AI ethics opinion say about competence?” and elsewhere states that “West Virginia’s opinion does not mandate client disclosure in all cases” — answers about the contents of an opinion the same page says does not exist, and a description of its consent posture that the Board’s actual opinion, with its written-consent requirement, contradicts.

Source page as published July 1, 2026 (archived copy: Wayback Machine). Contradicting primary source: Office of Lawyer Disciplinary Counsel, Legal Ethics Opinions page, accessed July 1, 2026.

4. Three claimed mandates that do not appear in the primary record

The Paxton post “2025 State Bar Guidance on Legal AI” (paxton.ai/post/2025-state-bar-guidance-on-legal-ai) presents three state requirements we could not verify against any primary source as of July 1, 2026.

Pennsylvania. The post states that Pennsylvania “mandates explicit disclosure of AI use in all court submissions,” calling it “a filing requirement” with “Implementation: August 2024.” What the primary record shows: Pennsylvania’s statewide instrument is the Pennsylvania Bar Association and Philadelphia Bar Association Joint Formal Opinion 2024-200 (May 2024; we link the most stable public copy), which is advisory by its own terms and is an ethics opinion, not a court-filing disclosure mandate. We could not locate any statewide Pennsylvania court-filing AI-disclosure mandate in the primary record as of July 1, 2026.

New York. The post states that New York “insists on at least two annual Continuing Legal Education (CLE) credits in practical AI competency,” with “Deadline: Q3 2025.” What the primary record shows: none of the New York instruments in the primary record — the New York State Bar Association’s Task Force on Artificial Intelligence report (April 2024), the New York City Bar’s formal opinions on generative AI, and the court system’s rule on AI use in court papers, 22 NYCRR Part 161 (effective June 1, 2026) — imposes an AI-specific CLE-credit requirement, and Part 161 itself is permissive and accountability-based, with no system-wide disclosure duty. We could not locate a New York AI CLE-credit requirement in any primary source as of July 1, 2026.

California. The post states that California “requires multi-jurisdictional compliance for AI cloud tools,” with “Rollout: January 2025.” What the primary record shows: California’s instruments are COPRAC’s Practical Guidance for the Use of Generative AI — guidance, first issued November 2023, with updated revisions approved by the State Bar Board of Trustees on May 14, 2026 (State Bar of California, Ethics & Technology Resources) — and proposed amendments to the Rules of Professional Conduct that had not been adopted as of our latest verification. Neither is a “multi-jurisdictional compliance” mandate with a January 2025 rollout, and we could not locate any such California requirement in the primary record as of July 1, 2026.

Source page as published July 1, 2026 (archived copy showing the same wording: Wayback Machine, May 9, 2026). Contrast sources: the Pennsylvania joint opinion, the NYSBA task force report, and the State Bar of California pages linked above, each accessed July 1, 2026.

5. “Colorado Bar Formal Ethics Opinion 145” as a March 2024 AI opinion: an apparently fabricated citation

The AI Vortex page at aivortex.io/legal/ethics/colorado/ (page dated April 11, 2026) states that “The Colorado Bar Association issued Formal Ethics Opinion 145 in March 2024, becoming one of the first state bars to publish a formal opinion specifically addressing AI use by attorneys,” labels it a “Formal Opinion,” and builds billing, confidentiality, and disclosure guidance on it, including the statement that “Opinion 145 has become a reference point for other state bars” — but the citation, as a March 2024 AI opinion, appears to be fabricated.

The Colorado Bar Association’s own published formal-opinions index does list a real Formal Opinion 145, but that opinion is “Discrimination; Bias,” dated May 14, 2022 — the number belongs to a 2022 opinion on a different subject — and because no opinion addressing artificial intelligence appears anywhere in the index, which ran through Formal Opinion 151 as of July 1, 2026, the March 2024 AI opinion the page describes appears to be fabricated (cobar.org/ethicsopinions). Colorado does have a real, verifiable AI instrument — Colorado Supreme Court Rule Change 2026(02), adopted January 8, 2026, which amended the Colorado Rules of Professional Conduct and is listed in the tracker with its primary source — but it is a supreme-court rule change, not a Colorado Bar Association ethics opinion, and as of our latest verification no CBA formal ethics opinion addresses AI. The same site’s Utah and Illinois pages are documented in item 1 and item 2 above.

Source page as published July 19, 2026 (archived copy: Wayback Machine). Absence check: Colorado Bar Association formal-opinions index, accessed July 1, 2026.

Corrections

If any item listed here has been corrected at the source, email info@mhsbsolutions.com — we verify and update within a week. The same standard applies to us: if anything on this page is wrong, send the primary source and we will fix it, date the fix on the changelog, and say what changed. Our own dataset and sourcing rules are documented on the methodology page.