Alaska: AI Ethics Guidance for Law Firms

Bottom line for a 5-50 attorney Alaska firm: Alaska Bar Association Ethics Opinion 2025-1 (Apr 2025) is the formal Alaska position on generative AI in legal practice and the controlling guidance for Alaska attorneys today. The opinion requires lawyers to understand the tools they use, obtain informed client consent before inputting client confidences into a generative AI tool, verify all AI-generated output, and align billing with existing fee rules. No Alaska state court or D. Alaska federal standing order requires AI disclosure as of 2026-04-23. Individual judges may impose requirements, and the opinion states attorneys must comply when courts require disclosure.

Start with: a written AI use policy, a citation verification log, and an informed-consent record for matters where client confidences may be input into a generative AI tool.

Alaska Bar Association Ethics Opinion 2025-1

Citation: Alaska Bar Association Ethics Opinion 2025-1, “Generative Artificial Intelligence & The Practice of Law,” adopted by the Alaska Bar Association Ethics Committee on 2025-04-03. Primary source PDF. Adopted opinions index.

Status: Advisory ethics opinion. The Ethics Committee’s opinions are not binding court rules but are the Alaska Bar’s formal statement of what the Alaska Rules of Professional Conduct require. The opinion is described as broadly similar in scope and content to ABA Formal Opinion 512 (July 2024), though ABA opinions are persuasive, not binding, in Alaska.

What it requires

The opinion sets seven binding requirements.

• Informed client consent is required, per ARPC 1.6, before inputting a client’s confidences or secrets into a generative AI tool. The lawyer must inform the client of the risk first.
• All AI-generated output must be carefully reviewed by the lawyer before use or submission to courts or clients, per ARPC 1.1 and ARPC 3.3 obligations.
• Lawyers must comply when a court requires disclosure of generative AI tool usage, per ARPC 3.3 candor obligations.
• Time saved through generative AI use cannot be billed to clients at the same rate as if the attorney had done the work manually. Time savings do not convert into billable time, per ARPC 1.5 fee obligations.
• Costs of generative AI tools and time spent training on those tools cannot be passed to clients absent client consent, per ARPC 1.5 fee obligations.
• When a firm intends to bill clients for the direct cost of AI tool use, it must disclose the client’s liability for those charges, along with the basis on which they will be calculated. The disclosure must arrive within a reasonable time after beginning the representation, per ARPC 1.5 fee obligations.
• Supervising lawyers who oversee junior lawyers or nonlawyer staff using generative AI must understand enough about the technology to provide appropriate oversight, per ARPC 5.1 and ARPC 5.3 obligations.

What it recommends

• Use AI tools tailored to law practice with contractual assurances of confidentiality, typically a “closed” system in which prompts are not used to train the tool and remain within the lawyer’s account.
• When using a tool that is not a closed system, anonymize prompts. The opinion notes anonymization may not be sufficient protection in all cases.
• Screen for bias in generative AI outputs, recognizing that AI models can produce biased results that may affect work product.
• Continue developing technological competency in generative AI as the technology evolves.

Rules cited

The opinion cites twelve ARPC rules. Competence, scope of representation, and diligence sit at ARPC 1.1, 1.2, and 1.3. Communication, fees, and confidential or secret client information sit at ARPC 1.4, 1.5, and 1.6. Meritorious claims at ARPC 3.1 and candor toward the tribunal at ARPC 3.3 govern court conduct. Truthfulness in statements at ARPC 4.1 and misconduct at ARPC 8.4 govern broader practice. Supervisory responsibilities of lawyers (ARPC 5.1) and over nonlawyers (ARPC 5.3) round out the list.

Technology competence baseline

Alaska adopted the technology-competence comment to ARPC 1.1 effective 2017-10-15. That pre-existing obligation is the foundation on which Opinion 2025-1 builds. Failing to adequately understand generative AI tools before using them in client matters is an independent ARPC 1.1 violation, separate from any errors the AI itself produces.

What federal courts in Alaska require for AI use in filings

No district-wide standing order requiring disclosure or certification of AI use has been identified for the U.S. District Court for the District of Alaska. Individual judges may have issued chamber-specific standing orders; counsel should check each assigned judge’s current standing orders before filing. The operative federal default is Fed. R. Civ. P. 11.

What Alaska state courts require for AI use in filings

No Alaska Supreme Court administrative order or court rule requiring AI disclosure in attorney filings has been identified at the statewide level. The Alaska Court System has deployed an internal AI-powered chatbot for public-facing probate assistance, but this is an administrative tool and does not constitute a standing order on attorney practice. Opinion 2025-1 states that lawyers must comply if courts require disclosure; in the absence of a standing order, the opinion does not independently require proactive disclosure of AI use to courts. ARPC 3.3’s candor obligation applies to all filings regardless of AI involvement.

What AI-related rules are pending in Alaska?

Senate Bill 2, 34th Legislature (2025-2026 Session), introduced 2025-01-22, addresses comprehensive AI governance. Coverage includes disclosure on AI-generated election communications, biennial agency AI inventories and impact assessments, citizen rights in agency AI decisions, and prohibitions on state-agency biometric identification, emotion recognition, and social scoring. As of 2026-04-23, SB 2 has been heard and held in Senate State Affairs and referred to Senate Judiciary; not enacted. If enacted, SB 2 primarily affects government-facing legal work and does not directly regulate private law firm AI use.

House Bill 47, 34th Legislature, criminalizes AI-generated child sexual abuse material, creates civil penalties for platforms enabling such generation, and addresses manipulated sexual imagery of adults and minor social media accounts. Passed the Alaska House on 2026-02-27; awaiting Senate consideration. It has low direct impact on AI tools used in legal practice but is relevant to criminal defense and family law practices.

House Concurrent Resolution 3 established a seven-member joint Legislative Task Force on Artificial Intelligence. The task force studies AI applications in state government, education, legal and ethical issues, economic opportunities, and governance frameworks. It expired 2026-02-02 after a required report deadline of 2026-01-31. The report may form the basis for future Alaska AI legislation.

How do malpractice carriers in Alaska treat AI use?

ALPS (Attorneys Liability Protection Society) is active in Alaska and is the carrier endorsed by more state bars than any other direct legal-professional-liability writer. Alaska does not mandate malpractice insurance as a condition of bar membership. Standard LPL policies generally do not contain explicit AI exclusions today, but coverage for AI-related claims is not guaranteed. Insurers may contest coverage under the “professional services” definition where AI-generated work product was submitted without adequate attorney review, or under intentional-act exclusions where an attorney deliberately relied on unverified AI output. Inputting confidential client information into public AI tools that harvest data for model training may implicate confidentiality-breach exclusions. AI-generated citation errors resulting in Rule 11 sanctions are generally not covered by LPL policies.

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

ALPS has published general guidance on AI risk but has not published Alaska-specific renewal-application items. Opinion 2025-1 establishes the substantive standard of care. Produce documentation that maps to it: the firm’s written AI use policy, vendor confidentiality review for each tool in use, informed-consent records for matters where client confidences are input, billing-disclosure language where AI costs are passed through, and the citation verification log. At renewal, request written confirmation from the carrier whether the current policy covers AI-related errors and omissions. Document the response.

What documentation should a Alaska firm keep on file?

Month one (foundational)

Three foundational artifacts establish the baseline.

1. (Owner: managing partner + firm administrator) Written AI use policy covering approved platforms (with completed data-privacy review), prohibited uses (inputting client confidences into non-closed public AI tools without consent), the mandatory review-before-filing requirement, and the supervision protocol for lawyers and nonlawyer staff. The policy maps to ARPC 1.1 and ARPC 5.1 and 5.3 obligations.
2. (Owner: litigation lead) Citation verification log documenting that all AI-generated citations and factual references are independently verified against primary sources before any court filing or submission to opposing counsel. The log responds directly to Opinion 2025-1’s candor holding under ARPC 3.3 obligations.
3. (Owner: matter lead attorney) Judge-specific standing-order check before filing in any Alaska state or federal proceeding, since no court-wide order exists but individual judge requirements may differ.

Months two and three (operational documentation)

Operational documentation extends the policy work into vendor and training records.

1. (Owner: firm administrator + outside IT) Vendor privacy review record for each AI platform in use: terms-of-service review, confirmation that the platform is a closed system (prompts not used for model training), data-retention policy, and contractual assurance of confidentiality. The record maps to ARPC 1.6 confidentiality obligations.
2. (Owner: firm administrator) Staff training records for all lawyers and nonlawyer staff using AI tools, including training on the firm’s AI use policy and on ARPC requirements. The records map to ARPC 5.1 and 5.3 supervisory obligations.

Months four to six (per-matter discipline)

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

1. (Owner: matter lead attorney) Client confidentiality consent documentation for any matter where client confidences or secrets will be inputted into a generative AI tool, obtained before the first such input. Recommended: add an AI-use disclosure provision to every engagement letter. The consent record maps to ARPC 1.6 confidentiality obligations.
2. (Owner: billing partner) Billing disclosure language for engagements where AI tool costs may be passed to clients: written disclosure of the client’s liability for those charges and the basis for calculating them, delivered within a reasonable time after engagement commencement. The disclosure maps to ARPC 1.5 fee obligations.
3. (Owner: managing partner) Carrier coverage review at each malpractice renewal cycle: written confirmation from carrier or broker whether the current policy covers AI-related errors, with any AI exclusions or riders documented.

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