October 18, 2026 (in 9 days): Tenth Circuit: comments close on 2027 rules, incl. proposed Rule 46.5(B)(4) generative-AI review →

Pennsylvania: AI Ethics Guidance for Law Firms

Formal opinion

Verified September 27, 2026

Citation
PBA/Philadelphia Bar Joint Formal Opinion 2024-200 (May 22, 2024)
Opinion date
May 2024

Summary

Pennsylvania has a joint formal advisory opinion (Joint Formal Opinion 2024-200, May 2024), a binding Pennsylvania Supreme Court Interim Policy governing judicial AI use (effective December 2025), and an active enforcement record including a Commonwealth Court brief-striking memorandum opinion (November 2025). Pennsylvania appears in our AI sanctions case tracker at one of the higher rates among US states; see the AI hallucination cases section below for the live count.

On this page
  1. Joint Formal Opinion 2024-200
  2. Pennsylvania Supreme Court Interim Policy on AI Use by Judicial Officers and Court Personnel
  3. What federal courts in Pennsylvania require for AI use in filings
  4. What Pennsylvania state courts require for AI use in filings
  5. Has anyone been sanctioned for AI use in Pennsylvania?
  6. What AI-related rules are pending in Pennsylvania?
  7. How do malpractice carriers in Pennsylvania treat AI use?
  8. What does my Pennsylvania malpractice carrier ask about AI at renewal?
  9. What documentation should a Pennsylvania firm keep on file?
  10. Court orders binding attorneys (11)
  11. Advisory and non-attorney rules (2)
  12. Pending AI cases (2)
  13. Decided AI hallucination cases (21)
  14. Applicable rules (reference)
This summary is informational only. Verify the primary source before relying on this entry in any filing or client matter. Bar rules differ meaningfully by state; consult a licensed attorney in your state.

Bottom line for a 5-50 attorney Pennsylvania firm: Joint Formal Opinion 2024-200 (May 22, 2024) is the governing standard for attorney AI use, paired with the Pennsylvania Supreme Court’s binding Interim Policy on judicial AI use (effective December 8, 2025) and an active enforcement record including a November 2025 Commonwealth Court memorandum opinion striking an attorney’s brief for AI-fabricated authority. The opinion’s citation-verification, confidentiality, and supervision obligations are concrete: every AI-generated citation must be independently verified against the primary source before filing. Pennsylvania appears repeatedly in the AI sanctions case tracker (live count below), making it one of the higher enforcement-volume states.

Start with: a written AI use policy keyed to Joint Formal Opinion 2024-200, a citation verification checklist for every AI-assisted filing, and a vendor confidentiality assessment for each AI tool the firm uses.

Joint Formal Opinion 2024-200

Citation: Joint Formal Opinion 2024-200 (adopted May 22, 2024; published June 2024), issued jointly by the Pennsylvania Bar Association Committee on Legal Ethics and Professional Responsibility and the Philadelphia Bar Association Professional Guidance Committee. Archived copy of the opinion PDF (Internet Archive capture of the PBA’s public listing, 2024-12-12; the PBA moved its formal opinions behind member login in its 2026 website migration, and the co-issuing Philadelphia Bar Association’s account of the opinion is the live official reference).

Status: Formal advisory opinion. The opinion expressly states it does not bind the Disciplinary Board of the Supreme Court of Pennsylvania or any court. Carries persuasive weight only.

What it requires

  • Competence (RPC 1.1, including Comment 8): Attorneys must be proficient in AI tools to the same degree as traditional tools. That includes understanding how generative AI differs from traditional AI, its benefits, risks, and failure modes including hallucination and training bias.
  • Citation verification (RPC 3.1, 3.3): All AI-generated citations must be independently verified for existence and accuracy against the primary source before filing or reliance. This is a concrete obligation, not aspirational.
  • Confidentiality (RPC 1.6): No confidential client information may be input into AI systems lacking adequate confidentiality and security protections. Attorneys must understand whether a tool retains, shares, or trains on submitted data.
  • Client communication and consent (RPC 1.4): Attorneys must explain how AI tools are being used and their potential impact on case outcomes. When AI use materially affects the representation, informed client consent must be obtained before using specific AI tools.
  • Truthfulness (RPC 8.4(c)): AI-generated content submitted to any tribunal or third party must be truthful, accurate, and based on sound legal reasoning. The attorney remains fully responsible.
  • Billing (RPC 1.5): AI-related costs must be reasonable. If AI tool charges are passed to clients they must be disclosed. AI efficiency savings should not be used to inflate fees.
  • Supervision (RPC 5.1, 5.3): Supervising attorneys and firms must establish policies ensuring all personnel comply with professional conduct rules when using AI. Delegating AI tasks to a paralegal does not relieve the supervising attorney of responsibility.

Notable gaps

The opinion does not address whether AI use must be disclosed to courts absent a specific court order. It does not specify which AI tools or categories are acceptable, and does not address fee arrangements for contingency matters when AI reduces time. It is advisory: a Pennsylvania disciplinary authority could interpret the underlying rules differently.

Aggregator fabrication flag

Multiple aggregator websites assert that Pennsylvania mandates explicit AI disclosure in all court submissions effective August 2024. That claim cannot be traced to any primary source order, rule, or statute, and is treated as fabricated until proven otherwise. The operative attorney standard is Joint Formal Opinion 2024-200, and no statewide disclosure mandate exists.

Pennsylvania Supreme Court Interim Policy on AI Use by Judicial Officers and Court Personnel

Citation: No. 643 Judicial Administration Docket; Pa.B. Doc. No. 25-1289; 55 Pa.B. (Sept. 20, 2025). Adopted September 9, 2025; effective December 8, 2025. Primary source PDF. Pennsylvania Bulletin entry.

Status: Binding interim administrative policy governing judicial officers and court personnel of the Unified Judicial System. Not a rule governing practicing attorneys, but it sets the confidentiality baseline the judiciary views as acceptable, and the precedent standard the courts will apply to attorneys by analogy. Non-public case information may not be shared with a “Non-Secured AI System” (those that do not guarantee confidentiality, non-retention, and no model training). Personnel “are responsible for the accuracy of their work”. The policy is explicitly interim; the AI Advisory Committee remains active and a permanent framework is anticipated.

On December 10, 2025, the Pennsylvania Judicial Ethics Advisory Board issued General Ethics Guidance No. 2-2025, Use of Generative Artificial Intelligence, addressed to judges and magisterial district judges rather than attorneys. It applies the judicial conduct codes to judicial use of generative AI under the Interim Policy, states that it is “incumbent upon Judicial Officers to check the accuracy and veracity of all work product created by GenAI,” and notes that no disclosure requirement for judicial use currently exists (tracker entry).

What federal courts in Pennsylvania require for AI use in filings

No district-wide AI standing order exists in any of Pennsylvania’s three federal districts; AI duties come from individual judges’ orders and practices. Judge Michael M. Baylson (E.D. Pa.) maintains an individual Standing Order Re: Artificial Intelligence (“AI”) in Cases Assigned to Judge Baylson (dated June 6, 2023). It requires that any attorney or pro se party who has used AI in preparing any complaint, answer, motion, brief, or other paper filed with the Court must, in a clear and plain factual statement, disclose that AI was used in any way in the preparation of the filing. The order also requires certification that each and every citation to the law or the record in the paper has been verified as accurate. Magistrate Judge José Raúl Arteaga (E.D. Pa.) sets out a more detailed version in section D of his practices and procedures (current version updated and effective September 17, 2026): an attorney or pro se party who uses generative AI to prepare a filing in a matter assigned to him must disclose that use, identify precisely which portions of the document contain generated content, identify the specific tool and how it was used, and certify that each citation to the law or the record was verified as accurate under Rule 11. Judge Kai N. Scott (E.D. Pa.) has two AI provisions (tracker entry). Her standing order (March 3, 2025) applies when generative AI is used in citing legal authority: the filer must disclose that use, name the AI program, and certify that each citation of legal authority has been verified. Section II.B.2 of her Policies and Procedures (edition marked “Effective Date: August 2026”; in place since the January 2025 edition) is broader: any attorney or pro se party who “has used Artificial Intelligence (‘AI’) in the preparation of any complaint, answer, motion, brief, or other paper” must “disclose that AI has been used in any way in the preparation of the filing, and CERTIFY, that each and every citation to the law or the record in the paper, has been verified as accurate.” Judge Kelley Hodge (E.D. Pa.) reminds anyone using generative AI in filings or discovery of existing Fed. R. Civ. P. 11(b) and 26(g) obligations (policies, section I.G.3) but has not issued a separate disclosure mandate. In the Middle District, Judge Karoline Mehalchick and Magistrate Judge Leo A. Latella require a Certificate of Use of Generative AI with any document prepared using a generative AI tool, disclosing the tool, the portions it prepared, and that a person checked the accuracy of the AI-generated content, including citations (Mehalchick order; Latella order). Magistrate Judge Phillip J. Caraballo’s chambers page, as read on 2026-09-28, cautions parties and counsel to verify any facts or legal citations prepared through generative AI and to safeguard confidential information, without a certificate requirement. The operative federal default in cases without an individual judge order is Fed. R. Civ. P. 11.

What Pennsylvania state courts require for AI use in filings

The Pennsylvania Supreme Court Interim Policy (above) governs judicial AI use, not attorney conduct. No statewide Pennsylvania rule mandates AI disclosure in attorney filings, but at least three county Courts of Common Pleas have entered administrative orders that do. Butler County’s order (In re: Use of Generative Artificial Intelligence, MsD No. 2024-40258, Sept. 10, 2024; 54 Pa.B. 5980) requires attorneys and pro se litigants to disclose any use of generative AI and to attach an affidavit attesting either that it was not used or that every citation was verified by a human being. Lancaster County (Docket No. CI-25-07564, Oct. 6, 2025) and Westmoreland County (Administrative Order No. 3 of 2026, May 8, 2026) require disclosure and an Affidavit of Use of Generative AI naming the tool, the AI-drafted portions, and a human accuracy check of citations. Firms filing in other counties may wish to check that county’s local rules and administrative orders. The active enforcement venue is the Commonwealth Court (see below). See the tracker entry for the Supreme Court Interim Policy.

Has anyone been sanctioned for AI use in Pennsylvania?

Pennsylvania has more documented AI hallucination incidents than nearly any other state. Spotlight PA documented at least 13 Pennsylvania cases in 2025 containing confirmed or suspected AI hallucinations. Three are concrete enough to anchor the firm-level analysis.

Commonwealth Court Memorandum Opinion No. 1172 C.D. 2025 (Pa. Commw. Ct. Nov. 24, 2025) (Cohn Jubelirer, P.J.) (unreported)

Primary source PDF. The Commonwealth Court struck an attorney-filed brief that “admittedly contains numerous factual and legal errors” attributable to unverified AI-generated content, and denied the motion to file an amended brief. The opinion describes, “[f]or informational and educational purposes,” Joint Formal Opinion 2024-200’s statement that attorneys must “check and verify all citations and the material cited”; it expressly declines to decide whether any ethical rule was violated. The case is tracked at Associated Builders & Contractors v. Bucks County Community College.

South Side Area School District v. Pennsylvania Human Relations Commission, No. 78 MD 2025 (Pa. Commw. Ct.)

Pending. After an en banc argument on December 10, 2025, Judge Matthew S. Wolf’s January 12, 2026 order listed, in a footnote, ten passages in petitioners’ brief whose quoted language “could not be located in the cited case,” including language attributed to Bayada Nurses, Inc. v. Department of Labor & Industry, 8 A.3d 866 (Pa. 2010). The order granted petitioners leave to file a letter addressing the issue and does not mention AI or sanctions. As of 2026-09-30 the court’s docket shows the case as awaiting consideration, with no ruling on the quotation issues. News coverage reported that the court raised AI at argument; that exchange is not in the court record, and the order notes that “oral argument is not transcribed.” See the tracker entry.

Saber v. Navy Federal Credit Union, No. 2449 EDA 2024 (Pa. Super. Oct. 28, 2025)

Unreported, non-precedential. A pro se appellant cited fabricated AI-generated cases; the court applied appellate waiver, treating arguments unsupported by valid authority as undeveloped and waived. First documented Pennsylvania Superior Court acknowledgment of the AI hallucination problem.

No Pennsylvania legislation specifically targeting attorney AI conduct is currently pending. Adjacent bills (HB 1533 on AI liability, HB 1857 on consumer AI disclosure) address related areas but not attorney practice directly. The Pennsylvania Supreme Court’s policy is designated interim, and its approving order sets no expiration; the AI Advisory Committee remains active.

How do malpractice carriers in Pennsylvania treat AI use?

Pennsylvania Insurance Department Notice 2024-04 (Pa.B. Doc. No. 24-484; 54 Pa.B. 1910; April 6, 2024) requires insurers operating in Pennsylvania to develop and maintain a written AI governance program. They must notify consumers when AI systems are used in decisions affecting them, conduct third-party vendor oversight and due diligence, and produce AI governance documentation during examinations. Malpractice carriers writing in Pennsylvania are subject to this notice and are themselves building AI governance frameworks. The supervision and verification obligations Joint Formal Opinion 2024-200 places on firms map directly onto the standard-of-care analysis a carrier or broker will use after a hallucination-driven claim. The Commonwealth Court’s November 2025 brief-striking memorandum opinion is now an explicit data point in that analysis.

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

No Pennsylvania-domiciled malpractice carrier has published renewal-application items specific to AI as of this entry’s date. The substantive standard the carrier will evaluate against is Joint Formal Opinion 2024-200 plus the Insurance Department’s Notice 2024-04 governance benchmarks. Firms should be ready to produce a written AI use policy, vendor confidentiality assessments, supervision and citation-verification protocols, and training records, with an incident log if any AI-assisted filing has produced an error. Firms with matters in front of Judge Baylson (E.D. Pa.) should also be ready to demonstrate a court-specific disclosure workflow. Confirm application items directly with the carrier or broker.

What documentation should a Pennsylvania firm keep on file?

Month one (foundational)

  1. (Owner: managing partner + firm administrator) Written AI use policy referencing Joint Formal Opinion 2024-200 as the governing standard. Identify approved tools, prohibit input of client confidential information into non-secured tools, require citation verification before any filing, and name the supervision chain for AI-assisted work product.
  2. (Owner: litigation lead) Citation verification checklist for every AI-assisted filing: supervising attorney certifies all citations verified against primary sources and all quoted material verified against original text. The Commonwealth Court’s November 2025 memorandum opinion is an in-state example for the cost of skipping this step.
  3. (Owner: matter lead attorney) Judge-specific AI order check at the opening of each federal matter. Include an explicit workflow for each chambers with an AI order: in E.D. Pa., Baylson (factual statement of any AI use plus citation-verification certification), Arteaga (disclosure of generative AI use, the generated portions, and the tool, plus citation certification), and Scott (under her Policies and Procedures, a factual statement of any AI use in preparing a filing plus certification of every citation to the law or the record; her standing order separately covers generative AI used to cite legal authority, naming the program); in M.D. Pa., Mehalchick and Latella (Certificate of Use of Generative AI). For state-court matters, the same check covers the county’s administrative orders: Butler, Lancaster, and Westmoreland Counties require AI disclosure with an affidavit (see the state-court section above).

Months two and three (operational documentation)

  1. (Owner: firm administrator + outside IT) Vendor confidentiality assessment for each AI tool: data retention policy, training/retention status, data processing agreement, benchmarked against the Pennsylvania Insurance Department Notice 2024-04 framework (written governance program, vendor due diligence).
  2. (Owner: managing partner + billing partner) Engagement letter provisions disclosing that the firm may use AI tools with appropriate safeguards. Include written client acknowledgment for sensitive matters. Addresses RPCs 1.4 and 1.6.
  3. (Owner: firm administrator) Attorney and staff training records showing each attorney and relevant staff member has reviewed Joint Formal Opinion 2024-200 and the firm’s AI policy. Required by RPCs 5.1 and 5.3.

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) Per-matter AI use notes recording which tools were used, what was input, and what review and verification steps were taken before filing.
  2. (Owner: managing partner) Incident log and periodic review capturing any instance where AI-generated output contained errors before filing and the corrective steps taken. Include a review schedule for new Pennsylvania Supreme Court rules, AI Advisory Committee output, and updated vendor terms.

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

Court orders binding Pennsylvania attorneys (11)

Federal and state court AI rules in force that impose an obligation on attorney filings in Pennsylvania: a disclosure, verification, or use restriction, or a rule that makes the filer answerable for accuracy.

Advisory guidance and non-attorney rules (2)

In force, but not an operative obligation on attorney filings: cautions and notices that restate existing duties, and rules that bind court personnel or parties rather than attorneys.

Pending AI cases in Pennsylvania (2)

These 2 cases are awaiting decision: a show-cause order, sanctions motion, or appeal is filed but the court has not yet ruled. Each gets a dedicated page now so it's already indexed when the ruling issues. Subscribe with the form below; the newsletter reports each ruling after it issues.

Decided AI hallucination cases in Pennsylvania (21)

Court, disciplinary and administrative matters addressing AI use in legal proceedings, including sanctions, warnings and responses that imposed no sanction; each case page states the outcome. Editorially flagged cases for Pennsylvania firms appear first with a "Why this matters" note; the remaining 18 entries collapse below.

Other Pennsylvania cases (18)

Applicable rules (reference)

How each rule applies to AI use in legal practice. Rule numbers reflect this state's own numbering where it diverges from the ABA Model Rules.

Rule 1.1 : Competence
In the firm AI policy, list approved tools and document what each one can and cannot do, where its data goes, and where it hallucinates. The competence duty rests on the lawyer, not the vendor.
Rule 1.4 : Communication
Decide once, by matter type, when AI use rises to a level the client should be told about. Document the threshold in the firm AI policy and reflect it in the engagement letter.
Pennsylvania note: Aggregator claims that PA mandates explicit AI disclosure in all court submissions effective August 2024 are fabricated; no primary source supports a statewide disclosure mandate. Joint Formal Op 2024-200 governs.
Rule 1.5 : Fees
Bill actual time, not pre-AI hourly time. If AI subscription costs are passed through to a client, disclose the basis in writing before the invoice issues.
Rule 1.6 : Confidentiality
Vet every AI tool that touches client data: data retention, training-data use, security posture, breach disclosure, and deletion on termination. No client-identifying input into a public-tier tool.
Rule 1.7 : Concurrent Conflicts of Interest
Conflict checks need to cover shared AI tools and vendor data pools, not just lawyer-level knowledge transfer.
Rule 3.1 : Meritorious Claims and Contentions
Verify legal basis before filing AI-drafted arguments. Hallucinated case theories are sanctionable under Rule 3.1 and FRCP Rule 11.
Rule 3.3 : Candor Toward the Tribunal
Personally read every cited case and verify every quoted authority before signing a filing. This is the rule behind every AI-hallucination sanction issued to date.
Pennsylvania note: Pennsylvania appears repeatedly in our AI sanctions case tracker. The Commonwealth Court struck a brief in November 2025 for AI-fabricated authority. See the live case list below.
Rule 5.1 : Responsibilities of Partners and Supervisory Lawyers
Adopt a written firm AI policy, require AI training at intake, and verify compliance. Recent sanctions decisions (Mata, Johnson v. Dunn) have credited firms with pre-existing written policies as a mitigating factor.
Pennsylvania note: PA Supreme Court Interim Policy (effective December 2025) governs judicial AI use, separate from attorney rules. Firm-level supervision under Joint Formal Op 2024-200 is the operative rule for lawyers.
Rule 5.3 : Nonlawyer Assistance
Treat AI tools the way the firm treats paralegals: written supervision protocol, attorney review before client or court delivery, and named accountability per matter.
Rule 5.5 : Unauthorized Practice of Law
Do not deploy AI tools that provide direct legal advice to non-lawyer end users. UPL exposure attaches to the lawyer or firm endorsing the tool.
Rule 8.4 : Misconduct
Submitting AI-fabricated authority, misrepresenting AI involvement, or weaponizing AI against opposing parties can constitute misconduct under 8.4(c) (dishonesty) or 8.4(d) (conduct prejudicial to administration of justice).