Legal AI Guidance Monitor — Aloha AI Consulting

Most collections of AI guidance are lists of links. A link tells you a document exists. It does not tell you whether the document binds you, what it actually decided, or — the part that gets lawyers into trouble — how far it can be stretched before it snaps. Every record below carries both halves.

JavaScript is off. The sweep-status banner above and the candidate list at the foot of this page are both filled in by a request to /api/data, so neither is showing you anything right now — ignore them. Everything that matters on this page is printed in the HTML: all 8 authorities, their analyses, the duty matrix and the convergence table are below, complete and in date order. Filtering and sorting are the only other things you lose.

Read and analysed

document opened before writing

Listed, not read

retrieval blocked · flagged

Binding authority

one published order · two standing orders

Candidates

—

official domain · unreviewed

How to read this register

Three kinds of document sit in this register and they carry entirely different weight. Getting them confused is the most common error in this area, and it runs in both directions.

Ethics guidance binds no one

ABA Formal Opinion 512, the Florida and D.C. opinions, and California’s Practical Guidance are advisory. They interpret rules of professional conduct; they do not enact them. Their force comes from being persuasive to the body that does discipline you — which is why the Ninth Circuit’s reach for California’s guidance in a sanctions order matters more than the guidance’s own disclaimer.

A standing order binds exactly one courtroom

The two Middle District of Florida orders in this register were entered two months apart, in the same district, by two judges, and they impose different things. One demands two certification texts reproduced verbatim under penalty of perjury; the other prescribes no wording at all. Neither is “the federal rule.” A filing protocol written for one of them fails the other, and that is the practical lesson: the compliance unit is the judge, not the district and not the country.

One document here is law

Lnu v. Blanche is a published Ninth Circuit order and therefore precedential in that circuit. It is also the document that locates the duty most precisely — not at research, not at drafting, but at signing and filing.

What “verified” means here, and what it does not

Verified means a person opened the document at the cited URL, read the passage the analysis rests on, and recorded the date. It does not mean the authority is current, that it has not been superseded, or that its analysis fits your matter. Where retrieval failed, the record says so and carries no summary — there is one such record below, and it is deliberately left unread rather than filled in from memory.

Authority boundary. Candidates discovered by the automated sweep are restricted to official domains but have not been opened by anyone. They are listed separately, below the register, and carry no characterisation. Commentary and commercial trackers are excluded entirely. Nothing here is legal advice.

Where the authorities agree

Independent bodies construing different rule sets have landed in the same place on four propositions. Convergence is the most useful signal a register like this can surface, because it tells a reader which positions are safe to plan around and which are still one jurisdiction’s view.

Time saved by a generative tool belongs to the client, not the lawyer.

Held or stated by

ABA · D.C. Bar

Note

Florida Opinion 24-1 reaches the same result through its prohibition on falsely inflated hours and on prorating unallocable subscription costs.

The lawyer must verify output; the tool is never the source of the duty being discharged.

Held or stated by

ABA · Florida · D.C. Bar · Ninth Circuit

Note

The Ninth Circuit converts this from an ethical duty into an enforced one.

Using generative AI is not itself misconduct.

Held or stated by

M.D. Florida · Ninth Circuit

Note

Judge Moe’s order states that no rule prohibits the use of artificial intelligence; the Ninth Circuit declined to sanction for use alone.

Client information may not enter a self-learning tool without informed consent.

Held or stated by

ABA · Florida

Note

California states the duty as a prohibition on inadequately protected tools rather than as a consent framework, which is a stricter default.

Which authority speaks to which duty

Read down a column to find every authority that addresses a duty you are worried about. A dot means the document treats that duty substantively, not that it merely mentions it. The New York opinion has no dots because it was not read.

Duties addressed by each authority in the register
Authority	Competence and technological understanding	Confidentiality and informed consent	Client communication	Candor to the tribunal	Supervision of lawyers, staff and tools	Fees, expenses and billing	Advertising and client intake	Disclosure or certification to the court
ABA — Formal ethics opinion	Addresses Competence and technological understanding	Addresses Confidentiality and informed consent	Addresses Client communication	Addresses Candor to the tribunal	Addresses Supervision of lawyers, staff and tools	Addresses Fees, expenses and billing	Does not address Advertising and client intake	Does not address Disclosure or certification to the court
California — State bar practical guidance	Addresses Competence and technological understanding	Addresses Confidentiality and informed consent	Does not address Client communication	Addresses Candor to the tribunal	Addresses Supervision of lawyers, staff and tools	Addresses Fees, expenses and billing	Does not address Advertising and client intake	Does not address Disclosure or certification to the court
Florida — Advisory ethics opinion	Addresses Competence and technological understanding	Addresses Confidentiality and informed consent	Does not address Client communication	Does not address Candor to the tribunal	Addresses Supervision of lawyers, staff and tools	Addresses Fees, expenses and billing	Addresses Advertising and client intake	Does not address Disclosure or certification to the court
NYC Bar — Formal ethics opinion	Does not address Competence and technological understanding	Does not address Confidentiality and informed consent	Does not address Client communication	Does not address Candor to the tribunal	Does not address Supervision of lawyers, staff and tools	Does not address Fees, expenses and billing	Does not address Advertising and client intake	Does not address Disclosure or certification to the court
D.C. Bar — Ethics opinion	Addresses Competence and technological understanding	Addresses Confidentiality and informed consent	Does not address Client communication	Addresses Candor to the tribunal	Addresses Supervision of lawyers, staff and tools	Addresses Fees, expenses and billing	Does not address Advertising and client intake	Does not address Disclosure or certification to the court
M.D. Florida — Judicial standing order	Addresses Competence and technological understanding	Does not address Confidentiality and informed consent	Does not address Client communication	Addresses Candor to the tribunal	Does not address Supervision of lawyers, staff and tools	Does not address Fees, expenses and billing	Does not address Advertising and client intake	Addresses Disclosure or certification to the court
M.D. Florida — Judicial standing order	Does not address Competence and technological understanding	Does not address Confidentiality and informed consent	Does not address Client communication	Addresses Candor to the tribunal	Addresses Supervision of lawyers, staff and tools	Does not address Fees, expenses and billing	Does not address Advertising and client intake	Addresses Disclosure or certification to the court
Ninth Circuit — Published sanctions order	Addresses Competence and technological understanding	Does not address Confidentiality and informed consent	Does not address Client communication	Addresses Candor to the tribunal	Addresses Supervision of lawyers, staff and tools	Does not address Fees, expenses and billing	Does not address Advertising and client intake	Addresses Disclosure or certification to the court

The register

ABA

ABA Formal Opinion 512 — Generative Artificial Intelligence Tools

ABA Comm. on Ethics & Prof’l Responsibility, Formal Op. 512 (July 29, 2024)

United States · ABA Model Rules · Formal ethics opinion · July 29, 2024

The ABA Standing Committee on Ethics and Professional Responsibility applies the Model Rules to generative AI, and reaches a confidentiality conclusion stricter than most lawyers expect.

Weight

Persuasive only. The Model Rules bind no lawyer until a state adopts them; Opinion 512 is the reference text most state opinions are measured against.

What it holds

Lawyers using generative AI must satisfy their duties of competence, confidentiality, client communication, supervision, meritorious contentions, candor to the tribunal, and reasonable fees. On confidentiality the Committee draws a hard line: because many self-learning GAI tools are built so that their output can lead directly or indirectly to disclosure of information relating to a representation, "a client’s informed consent is required prior to inputting information relating to the representation into such a GAI tool." Consent must be specific — the opinion says that adding general boiler-plate to an engagement letter purporting to authorise GAI use is not sufficient.

What it establishes

That the confidentiality analysis turns on the architecture of the tool, not on the lawyer's intentions. It also settles the billing question: a lawyer billing hourly must bill actual time expended, so time saved by a GAI tool belongs to the client, not the lawyer. Where a lawyer spends fifteen minutes prompting a tool and then reviews the draft, the opinion says the billable time is those fifteen minutes plus the review — not the hours the task would once have taken. The same reasoning is applied to flat and contingent fees: a flat fee unchanged after a tool makes the work dramatically faster may itself become unreasonable under Rule 1.5.

What it does not establish

Advisory. It binds no lawyer anywhere, and the Committee expressly ties its confidentiality conclusion to the risks and capabilities of GAI tools as they stood at publication in July 2024 — an unusual, and honest, self-limiting footnote. It says nothing about tools that do not receive client information, and it does not resolve whether a closed, non-self-learning, contractually walled deployment triggers the consent requirement at all.

Why it matters

Almost every state opinion issued since is written in this opinion's vocabulary, and the Ninth Circuit's sanctions order in Lnu v. Blanche reaches the same structural conclusion from the opposite direction: the violation is not the tool, it is the unverified filing. If you read only one document in this register, read this one — then read the one that governs your jurisdiction, because that is the one that binds you.

What to do with it

Compare your engagement letter against the boiler-plate warning. If your firm's AI consent language is a single clause added to a standard retainer, this opinion says it does not do the work you think it does. Separately, check whether your billing narrative can survive the "actual time expended" test.

Duties addressed

Competence and technological understanding
Confidentiality and informed consent
Client communication
Candor to the tribunal
Supervision of lawyers, staff and tools
Fees, expenses and billing

BAR

Practical Guidance for the Use of Generative AI in the Practice of Law — 2026 revision

State Bar of Cal. Standing Comm. on Prof’l Responsibility & Conduct, Practical Guidance for the Use of Generative Artificial Intelligence in the Practice of Law (rev. May 14, 2026)

California · State bar practical guidance · May 14, 2026

California was first to publish practical guidance in November 2023 and is the first to revise it for agentic systems that act without prompting.

Weight

Not a rule and not an ethics opinion. It is guidance the State Bar describes as guiding principles rather than best practices — but the Ninth Circuit cited it in Lnu v. Blanche when identifying the applicable standard for California-based lawyers.

What it holds

The Board of Trustees approved revisions on 14 May 2026 to guidance originally published on 16 November 2023. The revision adds an explicit treatment of agentic AI — systems that can autonomously perform tasks or workflows without human prompting. The guidance is direct about what autonomy does and does not change: agentic capability "does not substitute for a lawyer's independent professional judgment, nor does it alter a lawyer's existing ethical obligations." What it changes is the difficulty of supervision. Competence now requires understanding the nature and extent of a system's autonomy, and the guidance warns against configuring agentic systems so that the system makes substantive legal determinations.

What it establishes

That the profession's first-published guidance treats agentic AI as a supervision problem rather than a new category of permission. It also carries a flat confidentiality prohibition rather than a consent framework: a lawyer must not input confidential client information into any generative AI solution that lacks adequate confidentiality and security protections, must anonymise client information, and should read the terms of use to determine how inputs are used. The bias section notes that with multiple agents, underlying bias can compound.

What it does not establish

Guidance, not a rule — the State Bar says so on its face, and calls it a living document that will be revised again. It does not create a disciplinary standard of its own; it maps existing Rules of Professional Conduct and the State Bar Act onto a new tool. Separately, the State Bar has circulated proposed amendments to the Rules of Professional Conduct related to artificial intelligence for public comment, which means this guidance describes a position that is still moving.

Why it matters

California is the jurisdiction where the Ninth Circuit's disciplinary rules point for lawyers whose principal office is in the state, so this document is not merely advisory to a California litigator in federal court — it is the content the court reached for. It is also, at present, the only major guidance revised specifically for agentic tools, which is the direction firm deployments are moving.

What to do with it

If you have deployed anything that takes multi-step action on its own — drafting, filing, scheduling, contacting — the agentic section is the part to read. The operative question the guidance poses is not whether you supervised the output but whether you understood the extent of the system's autonomy before you switched it on.

Duties addressed

Competence and technological understanding
Confidentiality and informed consent
Supervision of lawyers, staff and tools
Fees, expenses and billing
Candor to the tribunal

BAR

Florida Bar Ethics Opinion 24-1

Fla. Bar Ethics Op. 24-1 (Jan. 19, 2024)

Florida · Advisory ethics opinion · January 19, 2024

The first state ethics opinion of the generative-AI era, and still the only one that works through client-intake chatbots as an advertising problem.

Weight

Advisory. The opinion states on its face that advisory ethics opinions are not binding.

What it holds

Lawyers may use generative AI but must protect confidentiality, remain responsible for their own work product, avoid improper billing, and comply with advertising restrictions. Three conclusions are unusually concrete. First, confidentiality turns on whether the tool is self-learning: the opinion recommends obtaining the affected client's informed consent before using a third-party tool if that use would disclose confidential information, and says consent is not required where no confidential information leaves the firm. Second, a lawyer may not delegate to generative AI any act constituting the practice of law, and must verify the accuracy and sufficiency of all research it performs. Third, a chatbot that communicates with prospective clients falls under the lawyer-advertising rules and must disclose that it is an AI program and not a lawyer or firm employee.

What it establishes

That the nonlawyer-assistant analogy does real work: the opinion reasons from Florida's existing opinions on cloud computing, overseas paralegals, device disposal and metadata, and concludes that a third party operating the tool does not relieve the lawyer of the supervisory duty. On costs it is stricter than most: a lawyer may charge the actual cost of using generative AI on a client's matter, must say so in advance and preferably in writing, and — where the actual per-matter cost cannot be determined — may not prorate the subscription but must treat it as overhead. Time spent developing minimal competence in the tool is not chargeable to the client.

What it does not establish

Advisory and Florida-specific; it construes the Rules Regulating The Florida Bar, not the Model Rules. It predates the agentic-tool question entirely, and the opinion itself says these concerns should not be treated as an exhaustive list. It makes no finding about any particular product.

Why it matters

It arrived six months before ABA 512 and reached the same place on billing by a different route. For any firm running website intake, the chatbot analysis is the most directly operational passage in this whole register: it converts a marketing decision into an advertising-rule compliance question, including the risk that a prospective-client relationship forms without the lawyer knowing.

What to do with it

If your website has a chat widget, check whether it identifies itself as an AI program, whether it carries a disclaimer limiting the firm's obligations, and whether it screens for people already represented by counsel. The opinion treats all three as the lawyer's responsibility, not the vendor's.

Duties addressed

Competence and technological understanding
Confidentiality and informed consent
Supervision of lawyers, staff and tools
Fees, expenses and billing
Advertising and client intake

BAR

Formal Opinion 2024-5 — Generative AI in the Practice of Law

N.Y.C. Bar Ass’n Comm. on Prof’l Ethics, Formal Op. 2024-5 (2024)

New York · Formal ethics opinion · August 7, 2024

Listed because it exists and is a primary source. Its contents are not characterised here, because this monitor could not retrieve it.

Weight

Advisory. A city bar association opinion construing the New York Rules of Professional Conduct; it binds no one and is not the New York State Bar.

Why there is no summary here

nycbar.org sits behind a CAPTCHA challenge that refuses automated retrieval (it answers HTTP 202 with a redirect to a challenge page), and the Internet Archive was unreachable during this review. Every other entry in this register was written after reading the document; this one was not, so no characterisation of its holdings appears here. It is retained rather than deleted because a reader looking for New York guidance should know the opinion exists — but a reader should open it themselves rather than rely on a summary this monitor cannot support.

Why it matters

New York practitioners have no state-level opinion of comparable prominence, so this is the document most often cited to them. That is precisely why an unread summary of it would be the most damaging thing on this page.

What to do with it

Open the link in a browser and read it directly. If you need a characterisation you can cite, this register is not it.

BAR

D.C. Bar Ethics Opinion 388 — Attorneys’ Use of Generative AI in Client Matters

D.C. Bar Legal Ethics Comm., Op. 388 (Apr. 2024)

District of Columbia · Ethics opinion · April 1, 2024

The opinion that states the verification standard in the plainest terms: general-audience generative AI is not a reliable substitute for cite-checking.

Weight

Advisory, construing the D.C. Rules of Professional Conduct.

What it holds

The opinion is confined to generative AI products, and implicates Rules 1.1, 1.2, 1.5, 1.6, 1.16, 3.3, 3.4, 5.1, 5.3 and 8.4. Before using a tool a lawyer must understand how it works and what it does, with regard to its dangers and its limitations, and must either have a reasonable basis for trusting its outputs or review and validate them. It states flatly that current general-audience generative AI is not a reliable substitute for traditional fact- and cite-checking. On confidentiality it directs lawyers to determine whether the provider will access or retain client information, warns that many products on the market are specifically designed to collect and use what users submit, and advises either finding a trustworthy alternative or restricting inputs to non-confidential material.

What it establishes

That the billing conclusion is not an ABA idiosyncrasy — D.C. reaches it independently: a lawyer can only bill for the time the lawyer actually spent, and under an hourly agreement the economies produced by the tool must inure to the client's benefit. Out-of-pocket GAI expenses may be billed where the client agreed to pay expenses. On supervision, Rules 5.1 and 5.3 require reasonable measures to ensure that supervised lawyers and nonlawyers use these tools consistently with the Rules, which the opinion suggests can include requiring staff to verify confidentiality protections before use.

What it does not establish

Advisory, and jurisdictionally narrow. The Committee expressly notes that the concerns it raises may be resolved or mooted for particular products in the future — so a conclusion resting on how a 2024-era product handled inputs does not automatically survive a change in that product. It does not address agentic systems.

Why it matters

D.C., the ABA and Florida agree on the billing rule and on the verification rule while construing three different rule sets. That convergence is the strongest signal in this register: a lawyer arguing that the position is unsettled has three independent bodies against them.

What to do with it

Treat the verification duty as non-delegable and document it. The three-jurisdiction convergence means "the tool said so" is not available as a defence in any of them.

Duties addressed

Competence and technological understanding
Confidentiality and informed consent
Supervision of lawyers, staff and tools
Fees, expenses and billing
Candor to the tribunal

COURT

Standing Order of Judge Moe Requiring Disclosure of the Use of Artificial Intelligence

Standing Order Requiring Disclosure of the Use of Artificial Intelligence, No. 6:24-cv-01987-AGM-RMN (M.D. Fla. May 27, 2026), ECF No. 193

U.S. District Court · Middle District of Florida · Judicial standing order · May 27, 2026

The strictest disclosure regime in this register: two permissible certifications, verbatim, under penalty of perjury, with doubt resolved in favour of disclosure.

Weight

Binding, and narrowly so. It is a standing order entered on the docket of one case before Judge Anne-Leigh Gaylord Moe. It governs filings in that case. It is not a district-wide local rule.

What it holds

Every filing must carry a certification stating whether artificial intelligence was used in its preparation. Only two certification texts are permitted, and they must be reproduced verbatim — the order states that only strict compliance will be recognised, because analysing the latent significance of "substantially compliant" certifications is an inefficient use of the court's finite resources. A filing without one of the two certifications may be stricken; if the filing is a motion, complaint or petition, the relief requested may be denied without prejudice. Where a filer is unsure whether something counts as artificial intelligence, the order directs that every doubt be resolved in favour of disclosing that it was used. Conventional Westlaw or LexisNexis research, and ordinary search engines, are expressly carved out and need not be disclosed.

What it establishes

That a court can impose a disclosure obligation without prohibiting anything: the order says plainly that no rule prohibits the use of artificial intelligence. Its operative mechanism is the affirmative obligation embedded in the certification itself — the filer certifies under penalty of perjury that they have read each case cited, and that failure to do so where a cited case does not exist or does not stand for the cited proposition may be treated as a Rule 11 violation.

What it does not establish

One judge, one case. It is not the practice of the Middle District of Florida, not the practice of the Orlando Division, and not a signal about any other judge — including Judge Berger in the same district, whose order takes a materially different form. Citing this order for the proposition that "federal courts require AI disclosure" would be a misuse of it.

Why it matters

It is the clearest available specimen of the verbatim-certification model, and the carve-out for traditional legal research is the sharpest line any court in this register has drawn around what "artificial intelligence" means for disclosure purposes. Given that Westlaw and Lexis both now embed generative features, that carve-out is also the provision most likely to be tested.

What to do with it

If you appear before Judge Moe, copy the certification text out of the order — do not paraphrase it, and do not carry over a certification drafted for another judge. If you do not, read it as a template for what a strict-compliance regime looks like before one lands on your docket.

Duties addressed

Disclosure or certification to the court
Candor to the tribunal
Competence and technological understanding

COURT

Standing Order of Judge Berger on Generative Artificial Intelligence

In re: Use of Generative Artificial Intelligence, Standing Order (M.D. Fla. Apr. 2, 2026) (Berger, J.)

U.S. District Court · Middle District of Florida · Judicial standing order · April 2, 2026

The same problem, a materially lighter instrument — a two-option certification with no prescribed wording and an explicit page-limit exemption.

Weight

Binding on all parties appearing before Judge Wendy W. Berger, through counsel or pro se. Entered in the Jacksonville Division as a general standing order rather than in a single case.

What it holds

All parties must certify at the end of any filing either that no generative AI was used in drafting and preparing it, or that where generative AI was used, any AI-drafted language — even if later edited by a human — was personally reviewed by the filer or another human for accuracy, and that all legal citations reference actual, non-fictitious authority and accurately reflect its contents. The certification does not count against page limits. Non-compliant filings may be stricken without substantive consideration and without leave to amend. Numerous failures to comply, or failures to properly review AI-generated content, will result in sanctions up to and including dismissal of claims with prejudice, default, and referral to the licensing authority. All filings remain subject to Rule 11, 28 U.S.C. § 1927, and applicable ethical rules.

What it establishes

That two judges of the same district, two months apart, chose different mechanisms for the same risk. Berger prescribes no wording and requires no perjury oath; the substantive requirement is human review of AI-drafted language and verification that every citation is real and accurately characterised. The escalation ladder is stated in the order itself rather than left to Rule 11.

What it does not establish

One judge. It is not a Middle District rule, and it does not govern filings before Judge Moe or any other judge. Its careful phrase "even if later edited by a human" resolves one ambiguity — editing does not launder the obligation — but the order defines neither "generative AI" nor the threshold of use that triggers option two, and it contains no carve-out for embedded research tools.

Why it matters

Read against the Moe order, this pair is the most useful thing in the register for anyone drafting a firm-wide filing protocol. A protocol built for one of these orders fails the other. The practical lesson is that per-judge checking, not a single national template, is the compliance unit.

What to do with it

Before filing in the Middle District of Florida, check the individual judge's standing orders — not the local rules, and not what you did last month in front of a different judge in the same courthouse.

Duties addressed

Disclosure or certification to the court
Candor to the tribunal
Supervision of lawyers, staff and tools

CASE

Lnu v. Blanche — Sanctions for Fabricated AI Authorities and Failures of Candor

Lnu v. Blanche, No. 24-4790 (9th Cir. June 3, 2026) (order) (Paez, Bea, and Forrest, JJ.)

U.S. Court of Appeals · Ninth Circuit · Published sanctions order · June 3, 2026

A published sanctions order that says, in terms, that the rules are not violated at the point of research and drafting but at the point of signing and filing.

Weight

Published, and therefore precedential within the Ninth Circuit. The only entry in this register that is law rather than guidance.

What it holds

Two attorneys filed briefs containing multiple nonexistent cases, misattributed quotations and gross misrepresentations of real cases, then represented the errors as typographical mistakes and denied that generative AI might have produced them. The panel imposed sanctions of $2,500 each, suspended both from practice before the court for six months, ordered them to serve the order on their clients, opposing counsel and the presiding judge in every pending matter and on every attorney at their firm, required all firm attorneys for two years to certify under penalty of perjury in every filing whether generative AI was used, naming the tool and certifying personal review of all citations and quotations, and directed the Clerk to serve the order on the State Bar of California.

What it establishes

The clearest available statement of where the duty attaches. The panel states that it does not sanction the attorneys for the simple fact that they or their subordinates used generative AI; the rules are not violated at the point of research and drafting, but at the point of signing and filing. If a brief contains cases or quotations that do not exist, it generally does not matter whether the hallucination came from an AI tool or from the lawyer's own natural intelligence. The order is equally clear about what drove the severity: had the attorneys disclosed in their Motion to Correct that AI had been used against firm policy and apologised for failing to check the brief, lesser sanctions may have been warranted. The gravity of the discipline was owed to the repeated failure of candor, not to the use of the tool.

What it does not establish

It is a disciplinary order under Federal Rule of Appellate Procedure 46(b) and Ninth Circuit Rule 46-2, applying the California Rules of Professional Conduct because the attorneys' principal office was in California. It does not create a duty to disclose AI use — the two-year certification requirement is a sanction imposed on these attorneys and their firm, not a rule of general application. It does not hold that using generative AI is misconduct, and the panel cites both the Ninth Circuit's own advisory committee note and California authority for the proposition that there is nothing inherently wrong with appropriate use.

Why it matters

Every guidance document in this register says verification is the lawyer's job. This is the document that shows the price. It also answers the question lawyers actually ask after a hallucinated citation is discovered — what do I do now — and the answer is the opposite of instinct: disclose the source of the error immediately to the court and opposing counsel. The panel says so expressly, and the six-month suspension is what the alternative cost.

What to do with it

Two things. First, if you discover a fabricated citation in something you filed, the order tells you to alert the court and opposing counsel at once and disclose the source of the error — including that it came from a generative tool. Second, note that a firm policy against AI use did not protect the firm here; it became an aggravating fact once the filing went out anyway.

Duties addressed

Candor to the tribunal
Competence and technological understanding
Supervision of lawyers, staff and tools
Disclosure or certification to the court

Discovered candidates — nobody has opened these

The weekly sweep searches official domains for new material. What it finds appears here without characterisation, because characterising a document nobody has read is exactly the failure this register exists to avoid. Treat these as leads.

Loading discovered candidates…

Discovered candidates are fetched at page load and cannot be shown without JavaScript. Nothing is being hidden from you: candidates are unreviewed search results by definition, and the verified register above does not depend on them.