TL;DR:
The District of Connecticut on November 4, 2026 entered an Agreed Supplement to the Standing Protective Order in Air-Evac EMS, Inc. et al v. Aetna Life Insurance Co. et al, Civil Action No. 3:26-cv-01021-SVN. The court expands the definition of Confidential information to cover proprietary data and protected health information, and it restricts input of Designated Material into generative AI tools to tightly controlled channels. Outputs derived from permissible AI use may be treated as attorney work product, while unauthorized disclosures to third-party AI platforms remain grounds for sanctions. The order impacts discovery workflows, data governance, and the ability to leverage AI in drafting and analysis. For trial teams, the ruling underscores the necessity of using secure, training-free or data-retaining-free AI environments and maintaining rigorous confidentiality practices throughout the litigation. This is a concrete, timely development that affects how involving AI in discovery is governed in federal practice.
What happened
- On November 4, 2026, the United States District Court for the District of Connecticut entered an Agreed Supplement to Standing Protective Order in Air-Evac EMS, Inc., Med-Trans Corp., Guardian Flight LLC, and Reach Air Medical Services LLC v. Aetna Life Insurance Co. and Aetna Health, Inc. (Civil Action No. 3:26-cv-01021-SVN). The order was published as a formal modification to the protective framework governing confidential discovery materials. The court’s action is memorialized in the official order, which supersedes prior terms through the conclusion of the litigation unless modified. (govinfo.gov)
- The supplement expands the scope of confidential material to include “proprietary technical, scientific, financial, business, health, or medical information” and, crucially, “protected health information” and “confidential health information.” This broadens protections for sensitive data that may arise in air-transport and medical-claims contexts. (govinfo.gov)
Key provisions affecting AI use in discovery
- The order adds a new paragraph titled “Use of Discovery Material in Conjunction with Generative Artificial Intelligence,” prohibiting input of Designated Material into generative AI tools or platforms except through (1) secure electronic discovery platforms with access limited to attorneys on the case and (2) enterprise generative AI tools that do not train on or retain submitted data and do not disclose data to third parties. Outputs from such permissible AI tools remain treated as attorney work product when they reflect or derive from Designated Material. (govinfo.gov)
- The protective order explicitly recognizes that outputs, prompts, and related logs or settings generated by AI tools can carry work product protection, but the rule does not eliminate disclosure obligations under Rule 26(a)(2) when required by the court or the case’s needs. (govinfo.gov)
- The modification also reinforces PHI protections, permitting PHI handling under the order with appropriate safeguards, while excluding sensitive nonparty information such as genetic testing data or other categories specified in the order. This creates an integrated framework for data privacy and discovery governance in AI-assisted workflows. (govinfo.gov)
- The order applies to all parties and, crucially, to “designated records” produced under federal rules or local rules, extending the protective regime beyond documents to the entire discovery process and related disclosures, including potential disclosures to non-parties and the court. (govinfo.gov)
Practical implications for trial teams
- Data governance becomes a courtroom issue. The order treats inputs to AI tools containing Designated Material as potential third-party disclosures, unless those inputs occur within secure, defined channels. Litigation teams must audit who has access to Designated Material and ensure that any AI usage conforms to the restricted pathways identified in the order. This includes vetting whether vendors’ tools are truly training-free and data-retention-free. (govinfo.gov)
- Attorney work product protection extends to AI outputs when they are derived from Designated Material. This creates a concrete incentive to centralize AI-assisted drafting and analysis within approved platforms, and to maintain careful logs of prompts and outputs that may influence trial strategy or witness preparation. Teams should coordinate with litigators and IT to ensure that AI-generated work product is properly preserved and identified. (govinfo.gov)
- PHI and sensitive information are protected more robustly in AI contexts. In matters involving health data, the order signals heightened caution and places a premium on secure handling, de-identification where possible, and strict compliance with HIPAA alongside the protective order. Trial teams must align AI workflows with these constraints to avoid inadvertent disclosures. (govinfo.gov)
Practical next steps for litigation teams
- Conduct an AI-use inventory. List all AI tools in use for document review, drafting, or analysis, and map them to the safeguards required by the Air-Evac order. Confirm that inputs containing Designated Material are routed only through approved secure platforms or enterprise tools that do not train or retain data. (govinfo.gov)
- Update internal AI governance policies. Create or revise an AI-use policy that (a) bans input of Designated Material into consumer AI services, (b) requires use of approved e-discovery platforms for confidential data, and (c) documents the provenance of AI-generated outputs as potential work product. Ensure that all litigation teams and outside counsel are trained on these requirements. (govinfo.gov)
- Preserve and document AI outputs. When using permissible AI tools, maintain records of the inputs and outputs that relate to Designated Material, so that counsel can assess whether the resulting materials remain protected as work product and are properly disclosed when required. (govinfo.gov)
- Coordinate with opposing counsel. Given that the order is a protective framework governing AI usage in discovery, ensure that any proposed deviations or requests for access to AI-assisted work product are discussed in light of the protective order and applicable rules. The case demonstrates that courts may tailor AI governance alongside traditional discovery disputes. (govinfo.gov)
Why Objection Academy remains relevant in light of AI governance
Even as courts impose tighter controls on AI in discovery, the core litigation skills of articulating objections, cross-examining witnesses, and applying the Federal Rules of Evidence remain critical. Objection Academy provides practical drills for handling objections and preserving trial-readiness, which is especially valuable as teams navigate AI-enabled workflows and the potential implications for evidentiary objections in real time. The program’s emphasis on objection technique, evidence application, and courtroom readiness aligns with the heightened need for disciplined advocacy when AI-generated materials and protective orders shape how evidence is presented and challenged at trial. Integrating strong objection training with compliant AI practices can help trial teams maintain credibility and control in the courtroom, even as technology reshapes discovery and preparation.
Conclusion
The Air-Evac EMS v. Aetna Life Insurance order marks a clear and timely milestone in the governance of generative AI within federal discovery. By expanding the scope of confidential information and by restricting input to AI tools while preserving work product protections for AI outputs, the District of Connecticut provides a concrete framework that trial teams can implement immediately. The ruling emphasizes the necessity of secure, accountable AI environments and disciplined data governance in modern litigation. As AI usage continues to evolve, defendants and plaintiffs alike should align their discovery workflows with these protections to minimize risk and preserve strategic advantages in trial.
Sources
United States District Court for the District of Connecticut, Air-Evac EMS, Inc. et al v. Aetna Life Insurance Company et al, Agreed Supplement to Standing Protective Order, Civil Action No. 3:26-cv-01021-SVN, November 4, 2026. GovInfo PDF: USCOURTS-ctd-3_26-cv-01021-0.pdf. (govinfo.gov)
Case Law Wire, Air-Evac EMS, Inc. et al v. Aetna Life Insurance Company et al, Court restricts use of confidential discovery material in generative AI tools, November 4, 2026. (caselawwire.com)
California Courts Newsroom, Use AI New Rules and Maybe a Law Are Horizon for California Lawyers, September 3, 2026 (context on GenAI rules that may affect practitioner behavior). (newsroom.courts.ca.gov)
California Bar practical guidance, Generative AI Practical Guidance for Attorneys, 2026 Update (supporting the broader context of AI ethics and practice). (calbar.ca.gov)