Colorado: AI Ethics Guidance for Law Firms

Bottom line for a 5-50 attorney Colorado firm: No formal CBA ethics opinion on AI exists, but Rule Change 2026(02) (effective January 8, 2026) states that technology’s role “does not diminish a lawyer’s responsibilities” under the Rules, and People v. Crabill (suspension of one year and one day, 90 days served) shows the discipline that follows filing unverified AI citations. Several D. Colo. judges require AI certifications: Judge Wang on every filing, Judge Kane on any filing prepared in whole or in part with generative AI, and Judge Crews on substantive motions and their briefing. Colorado’s automated decision-making technology law (SB 26-189) applies to consequential decisions made on or after January 1, 2027; it does not list legal services as a covered domain, but it can reach a firm’s own use of such tools in hiring and other employment decisions. Firms documenting AI governance may wish to consider a verification protocol, a vendor confidentiality review, and an inventory of any AI tool used in employment decisions.

Start with: a written AI use policy, a citation verification log for court filings, and vendor due-diligence records for each AI tool.

Colorado Bar Association practice guidance (no formal opinion)

Citation: Maria E. Berkenkotter and Lino S. Lipinsky de Orlov, Artificial Intelligence and Professional Conduct, 53 Colo. Law. 20 (Jan./Feb. 2024). Primary source.

Status: Practice commentary in Colorado Lawyer, the CBA’s official publication, co-authored by a sitting Colorado Supreme Court Justice and a Colorado Court of Appeals Judge. Not a formal CBA Ethics Committee opinion. The CBA’s formal ethics opinion index, through Opinion 151 (restrictive covenants, adopted November 8, 2025), contains no AI-specific opinion.

What it flags

• The article poses questions under Colo. RPC 1.1, 1.4, 1.5, 1.6, 3.3, 5.1, 5.3, and 8.4(c), (d), and (g), and under the Code of Judicial Conduct.
• Firms documenting AI governance may wish to note that the article appeared in the CBA’s official publication under judicial authorship.

Notable gap

The article is educational, not prescriptive. It does not define what an “adequate” verification protocol looks like.

Rule Change 2026(02): Scope and RPC 1.1 comments

Citation: Colorado Supreme Court, Rule Change 2026(02), “Amended and Adopted by the Court, En Banc, January 8, 2026, effective immediately.”

Status: Supreme Court amendment, adopted en banc, to the Scope section and the Rule 1.1 comments. The text of Rule 1.1 itself is unchanged.

What it does

• New Scope paragraph [20A]: “Technology, including artificial intelligence and similar innovations, plays an increasing role in the practice of law, but that role does not diminish a lawyer’s responsibilities under these Rules. A lawyer who uses, directly or indirectly, technology in performing or delivering legal services may be subject to discipline for a resulting violation of these Rules.”
• New Comment [9] to Rule 1.1 lists Rules that a lawyer’s use of technology, “particularly artificial intelligence,” can implicate: communication (1.4), reasonable fees (1.5), confidentiality (1.6), meritorious claims and defenses (3.1), candor toward the tribunal (3.3), supervisory and subordinate lawyers (5.1, 5.2), nonlawyer assistance (5.3), communications about a lawyer’s services (7.1), and bias (8.4(g)).
• Revised Comment [8] states that a lawyer should keep abreast of the benefits and risks associated with relevant technology.

What it does not do

• It does not create new affirmative disclosure or certification obligations.
• It does not define what competent AI use looks like operationally.

What federal courts in Colorado require for AI use in filings

No district-wide AI rule has been adopted in D. Colo.; the local rules contain no AI provision, and the operative default is Fed. R. Civ. P. 11. Several judges have their own requirements:

• Judge Nina Y. Wang: her standing order (effective December 1, 2025) provides that “[e]very filing shall contain an AI Certification regarding the use, or non-use, of generative AI,” signed by all individuals who contributed to drafting, and covers AI-drafted language even if later edited by a human. In a minute order in Hessert v. The Street Dog Coalition, No. 1:25-cv-03807 (D. Colo. Apr. 17, 2026), Judge Wang rejected a pro se plaintiff’s First Amendment, due process, equal protection, work-product, and separation-of-powers objections to the order as “without merit.” See the tracker entry.
• Judge S. Kato Crews: his Standing Order for Civil Cases (revised January 1, 2026) requires that “every substantive motion,” including motions under Rules 12, 56, and 65, “and the corresponding response and reply, shall contain an AI Certification.” The preparer certifies either that no portion was drafted by AI or that any AI-drafted language was personally reviewed by a human for accuracy and all citations are to actual authority. Filings without it are stricken without prejudice, and the certification does not count against page limits. See the tracker entry.
• Judge John L. Kane: his Pretrial and Trial Procedures memorandum for civil cases (revised March 2026) requires that any filing prepared in whole or in part using GenAI include a certification identifying the portion prepared with GenAI and affirming that any GenAI-drafted language was personally reviewed for accuracy using print reporters or traditional legal databases. See the tracker entry.
• Magistrate Judge Susan Prose: her standing order (October 21, 2024) requires an AI certification on specified motions, and in referral cases only on nondispositive opposed motions. See the tracker entry.
• Magistrate Judge Maritza Dominguez Braswell: her AI guidance states that it “is not an order and does not establish new rules or requirements”; it sets out a confer-first framework for addressing hallucinations found in an opponent’s filing (tracker entry). On the discovery side, in Morgan v. V2X, Inc., No. 1:25-cv-01991 (D. Colo. Mar. 30, 2026) (order; docket), she granted in part the defendant’s motion to amend the protective order. The court held that a litigant’s use of AI to prepare for litigation is protected work product under FRCP 26(b)(3), but that the name of any AI tool used with confidential information must be disclosed. The amended order bars uploading CONFIDENTIAL discovery to any AI platform unless the provider is contractually prohibited from storing or using inputs to train its model and from disclosing inputs to third parties, and the user can delete the confidential information. The plaintiff was ordered to disclose any platform used with confidential material within ten days. The order may serve as a model for AI-specific protective-order provisions. See the tracker entry.

Firms documenting compliance may wish to check the assigned judge’s standing orders at the opening of each federal matter.

What Colorado state courts require for AI use in filings

Colorado has no state-court attorney AI disclosure rule. Rule Change 2026(02) is a Rules of Professional Conduct amendment, not a court-filing rule. Colorado’s appellate courts have warned that AI-hallucinated citations are sanctionable. In Al-Hamim v. Star Hearthstone, LLC, 2024 COA 128 (Dec. 26, 2024), the Court of Appeals held that the self-represented appellant’s brief containing hallucinations violated C.A.R. 28(a)(7)(B). It declined to sanction him but warned him, “as well as lawyers and self-represented parties who appear in this court, that we will not ‘look kindly on similar infractions in the future,’” and stated that a future filing containing GAI-generated hallucinations “may result in sanctions.”

People v. Crabill, 23PDJ067 (Colo. O.P.D.J. 2023-11-22)

Court and judge: Colorado Office of the Presiding Disciplinary Judge. PDJ summary of the approved stipulation to discipline (PDF).

Facts: Attorney Zachariah Crabill used ChatGPT to generate case citations and submitted them in a motion without reading the cases. Before a hearing, he discovered they were fictitious. He did not alert the court or withdraw the motion. When confronted, he falsely attributed the errors to a legal intern. Six days after the hearing he filed an affidavit disclosing his ChatGPT use.

Findings: Through this conduct, Crabill was found to have violated Colo. RPC 1.1 (competence), 1.3 (diligence), 3.3(a)(1) (candor toward the tribunal), and 8.4(c) (dishonesty, fraud, deceit, or misrepresentation).

Discipline: Suspension for one year and one day, with ninety days to be served and the remainder stayed upon successful completion of a two-year period of probation, with conditions.

What AI-related rules are pending or newly enacted in Colorado?

Colorado automated decision-making technology law (SB 26-189). SB 24-205, the Colorado AI Act (signed May 17, 2024), never took effect. SB 25B-004 postponed its requirements to June 30, 2026, and SB 26-189, signed May 14, 2026, repealed and reenacted C.R.S. title 6, article 1, part 17 as “Automated Decision-Making Technology in Consequential Decisions” before that date arrived. The reenacted part “takes effect January 1, 2027” and “applies to consequential decisions made on or after January 1, 2027” (signed act, section 5). Key features:

• Scope. It regulates developers and deployers of automated decision-making technology (ADMT) used to materially influence a consequential decision about a consumer. Legal services is not a covered domain; the covered domains are education, employment, residential real estate, financial or lending services, insurance, health-care services, and essential government services and public benefits (6-1-1701(6)). Because “consumer” includes an employee and a Colorado-resident job applicant, a firm that uses covered ADMT to materially influence hiring or other employment decisions may be a deployer. A tool used “solely to summarize, organize, translate, draft, route, or present information for human review” is excluded from the ADMT definition.
• Deployer duties. Clear and conspicuous notice before using covered ADMT to materially influence a consequential decision (6-1-1704(1)). After an adverse outcome, a plain-language description of the decision and the ADMT’s role within thirty days (6-1-1704(3)), an opportunity to correct inaccurate personal data, and, to the extent commercially reasonable, meaningful human review and reconsideration (6-1-1705(1)). Records must be kept for three years (6-1-1703).
• What it dropped. The reenacted part has no impact-assessment, risk-management-program, or small-deployer (under-50-employee) provisions, and it provides that “compliance with the requirements of this part 17 is not a defense” to noncompliance with other applicable law (6-1-1707(8)).
• Enforcement. The Attorney General enforces it exclusively, as a deceptive trade practice under the Colorado Consumer Protection Act, with a sixty-day notice-and-cure process where a cure is possible (that cure provision is repealed January 1, 2030). It creates no new private right of action (6-1-1706, 6-1-1709).
• Rulemaking. The Attorney General filed proposed rules (4 CCR 904-6) on August 11, 2026, taking written comments between August 11 and October 26, 2026, with a public hearing on October 26, 2026 at 10:00 AM (coag.gov/ai; notice of hearing). The same rulemaking covers the Chatbot Safety Act (HB 26-1263), which the Attorney General’s page says takes effect January 1, 2027. See the framework page for more detail.

State-court civil filings. The Colorado Supreme Court’s Civil Rules Committee placed an AI item on its September 25, 2026 agenda: a county judge’s request for help with AI-generated filings, suggesting a rule requiring parties to verify that a pleading is correct (agenda packet). No rule has been proposed or adopted.

How do malpractice carriers in Colorado treat AI use?

No Colorado malpractice carrier has published explicit AI-specific underwriting guidance in materials reviewed for this entry. The substantive standard carriers point to is Rule Change 2026(02) plus the Crabill discipline order. The Crabill coverup element is particularly relevant to insurance posture. A false attribution or denial of AI use later in a malpractice proceeding may implicate intentional misconduct exclusions.

What does my Colorado malpractice carrier ask about AI at renewal?

No Colorado malpractice carrier has published AI-specific application items in materials reviewed for this entry. The substantive standards the carrier or broker references are Rule Change 2026(02), Crabill, the D. Colo. certification orders, and, for any firm using AI tools in employment decisions, SB 26-189. Plan on producing six items. First, a written AI use policy. Second, citation verification logs. Third, vendor due-diligence records, including data retention and training restrictions. Fourth, a federal-court certification compliance procedure. Fifth, an inventory of any AI tool used to materially influence employment or other consequential decisions, with the notice and adverse-decision procedures SB 26-189 requires. Sixth, a supervision protocol for paralegal and contract-attorney AI use. SB 26-189 applies to consequential decisions made on or after January 1, 2027, regardless of renewal cycle.

What documentation should a Colorado firm keep on file?

Month one (foundational)

Three foundational items.

1. (Owner: managing partner + firm administrator) Written AI use policy covering confidentiality vetting, citation verification, billing practices, and supervision. Anchored to Rule Change 2026(02) and the Rules its Comment [9] lists.
2. (Owner: litigation lead) Citation verification log or checklist for every court filing. Crabill shows the discipline that follows unverified AI citations, and Al-Hamim warns of future appellate sanctions.
3. (Owner: matter lead attorney) Federal-court AI certification check at the opening of each D. Colo. matter. Confirm whether the assigned judge (currently including Wang, Crews, Kane, and Prose) requires an AI certification. Log AI use or non-use accurately during drafting.

Months two and three (operational documentation)

Three documentation items.

1. (Owner: firm administrator + outside IT) Vendor due-diligence records for each AI tool. Document whether the vendor agreement prohibits training on client data and retention of uploaded materials. Confirm current data processing terms. Anchored to RPC 1.6 and Rule Change 2026(02).
2. (Owner: managing partner) Supervision protocol covering associate, paralegal, and contract-attorney AI use under RPCs 5.1 and 5.3. Identify approved tools, prohibited uses, and required attorney review of AI-generated work product.
3. (Owner: firm administrator) ADMT inventory listing every AI tool the firm uses in hiring, compensation, or other employment decisions, and any other decision within SB 26-189’s covered domains. For each, record whether it materially influences the decision and, if so, the notice, adverse-decision explanation, correction, and human-review procedures in place by January 1, 2027.

Months four to six (per-matter discipline)

These are recurring practices, not one-time projects. Each new matter exercises them.

1. (Owner: managing partner + billing partner) Engagement letter provision disclosing the firm’s AI use, identifying categories of use, and confirming confidentiality safeguards. Anchored to RPCs 1.4 and 1.6.
2. (Owner: matter lead attorney) Per-matter AI use notes capturing the tool used, the prompt category, the reviewer who verified output, and any client confidentiality decision. Anchored to Crabill, Rule Change 2026(02), and the D. Colo. certification orders.

We are building practitioner resources for firms working through this list. The newsletter covers new resources as they ship.