Connecticut Court Rules for Artificial Intelligence: Pleading Standards in 2026

Connecticut court rules for artificial intelligence establish that any practitioner or self-represented litigant who utilizes machine-learning tools to draft or edit court pleadings must independently verify all generated citations, legal authorities, and factual evidence before filing. Enacted statewide under Connecticut Practice Book § 4-9 and reinforced through the signature certifications of § 4-2(b), this codification replaces piecemeal local chamber orders with a unified procedural mandate, providing that unverified AI output may result in severe judicial sanctions, including the immediate entry of a nonsuit or default judgment.

How Did Connecticut Codify Court Rules for Artificial Intelligence?

For years, the American judicial landscape addressed generative artificial intelligence through an ad-hoc mosaic of individual judge standing orders. While individual federal and state trial judges issued disparate declarations—ranging from total prohibitions to mandatory certificates of disclosure—statewide procedural uniformity remained absent.

On June 11, 2026, the Judges of the Connecticut Superior Court adopted a comprehensive regulatory package that fundamentally shifted this dynamic. Rather than waiting for the standard annual Practice Book effective date of January 1, 2027, the Superior Court Judges ordered that the new Connecticut Practice Book § 4-9 (“Generative Artificial Intelligence (‘Generative AI’) Compliance”) and accompanying amendments to § 4-2(b) take immediate effect upon official publication on June 23, 2026.

The result is one of the nation's first binding, statewide procedural frameworks that directly integrates generative AI compliance into the foundational act of signing and filing a legal document.

What Does Connecticut Practice Book § 4-9 Specifically Require?

The text of Connecticut Practice Book § 4-9 establishes four explicit operational standards that apply across all civil, family, criminal, and housing matters filed in the Connecticut Superior Court:

1. Broad Statutory Scope (§ 4-9(a))

The rule defines generative AI comprehensively:

"any machine-learning model/application that can create original content—such as text, images, video, audio or software code—in response to a prompt or request, including but not limited to, ChatGPT, Co-Counsel, Google Bard, Neeva, Harvey, Ironclad, Bing, DeepSeek, Grok, and similar technology."

By naming both enterprise legal workflows and consumer platforms while appending a flexible technological standard, the rule prevents litigants from arguing that specialized legal tools escape oversight.

2. Mandatory Independent Verification and Case-Dispositive Sanctions (§ 4-9(b))

Addressing the documented risk of hallucinated citations and fabricated quotes, subsection (b) enacts a mandatory duty of human review:

"Due to the risk that generative AI can create inaccurate factual and legal information, including, without limitation, faulty citations to legal authority, fabricated quotations from such authority, and inaccurate or fabricated evidence, any person who uses generative AI in the creation or editing of any document filed with the court shall independently verify all citations, legal authorities or evidence produced by generative AI."

Crucially, the rule couples this duty with immediate case-dispositive consequences:

"The failure to do so may result in court-imposed sanctions, including, without limitation, the entry of a nonsuit or default judgment."

Where traditional sanctions under general procedural rules often required protracted motions for contempt or formal Rule 11 hearings, § 4-9(b) explicitly arms judges with the authority to throw out a plaintiff's action via nonsuit or enter an immediate default judgment against a defending party for unverified AI submissions.

3. Affirmative Representation upon Filing (§ 4-9(c))

Subsection (c) operates as an automatic legal warranty. Anyone submitting a document to the court represents that they have reviewed § 4-9 and made good-faith, diligent efforts to ensure full compliance with the Practice Book, the Connecticut Rules of Professional Conduct, and state law.

4. Sole Responsibility on the Filer (§ 4-9(d))

The court and its clerks are explicitly relieved of policing filings. Subsection (d) dictates that compliance rests solely with the filer. Litigants cannot shift blame to automated e-filing screening or administrative acceptance by the clerk's office.

How Does Revised Practice Book § 4-2(b) Enforce Compliance?

The procedural teeth of § 4-9 are anchored in Connecticut Practice Book § 4-2 (“Signing of Pleadings”). Effective June 23, 2026, the Superior Court Judges revised subsection (b) to expand the legal certification executed by an attorney or pro se party when signing any pleading, motion, objection, or request.

Under revised § 4-2(b), a signature certifies that:

1. The signer has read the document;
2. To the best of their knowledge, information, and belief, there is good ground to support it;
3. It is not interposed for delay;
4. The filer has complied with § 4-7 regarding the exclusion or redaction of personal identifying information (PII); and
5. The filer has complied with § 4-9 regarding generative AI.

By integrating § 4-9 into the core signature certification, Connecticut eliminates the need for separate, redundant standalone AI affidavits on every docket entry. Instead, every signature on a court document operates as an enforceable legal warranty that all AI-generated citations, quotes, and factual records have been independently verified against primary authorities.

What Are the Operational Differences: Verification vs. Disclosure?

A key debate during the drafting of AI court rules centered on whether to require affirmative disclosure (filing a "Certificate of AI Use") versus mandatory verification.

Connecticut's framework decisively favors mandatory independent verification:

• No Blanket Disclosure Mandate: Attorneys are not required to confess the use of assistive technology simply because an AI system summarized deposition testimony or generated initial prose drafts.
• Strict Liability for Verification: The court focuses squarely on substantive integrity. If a brief is filed with fictitious citations or distorted statutory quotations, the attorney is subject to disciplinary action under Rule 3.3 (Candor toward the Tribunal), Rule 1.1 (Competence), Rule 5.3 (Supervision of Nonlawyer Assistance), and § 4-9 sanctions.
• Equal Burden on Pro Se Litigants: Unlike certain federal jurisdictions that exempted unrepresented parties from complex local standing orders, Connecticut § 4-9 expressly applies to "any person who uses generative AI in the creation or editing of any document filed with the court." Self-represented litigants who rely on commercial chatbots to write motions face the exact same nonsuit and default judgment standards.

For further exploration of how judicial verification intersects with federal practice, examine our analysis of AI Court Standing Orders and Judicial Certification Rules, our breakdown of Rule 11 AI Verification Doctrine, and our review of AI Judicial Evidence Rules and Deepfake Admissibility.

Disclaimer: This analysis is provided for general informational and educational purposes only and does not constitute formal legal advice. Litigants and practitioners must review the official Connecticut Practice Book and specific court calendars for local practice guidance.