Lead story
ACEDS Analysis Argues Elusion-Only Recall Estimation Overstates TAR 2.0 Performance
ACEDS published analysis arguing that relying solely on elusion sampling to estimate TAR 2.0 recall systematically overstates performance, and that AI may better enable requesting parties to demonstrate production inadequacies.
- What happened
- ACEDS published analysis arguing that relying solely on elusion sampling to estimate TAR 2.0 recall systematically overstates performance, and that AI may better enable requesting parties to demonstrate production inadequacies.
- Why it matters
- Practitioners relying exclusively on elusion sampling for TAR 2.0 recall validation may face increased challenge from opposing parties. The analysis reinforces that defensible production requires multi-stratum sampling and transparent methodology documentation.
- Watch
- Analysis published June 26, 2026. Author is affiliated with ACEDS.
SourcesJD Supra
Briefing item
Federal Magistrate Orders Production of Expert's ChatGPT Prompts in Climate Litigation
A federal magistrate judge in Connecticut ordered production of AI prompts used by testifying expert Dr. Naomi Oreskes, who relied on ChatGPT to sort documents in Shell climate litigation, reasoning the prompts were part of the expert's methodology.
- What happened
- A federal magistrate judge in Connecticut ordered production of AI prompts used by testifying expert Dr. Naomi Oreskes, who relied on ChatGPT to sort documents in Shell climate litigation, reasoning the prompts were part of the expert's methodology.
- Why it matters
- This ruling signals that courts may increasingly treat generative AI use by experts as discoverable methodology rather than protected research process. Litigation teams using AI tools for document review or analysis should document their processes and anticipate disclosure obligations.
- Watch
- Published June 26, 2026. Order is being challenged; final disposition pending.
SourcesJD Supra
Briefing item
Court Holds Email Search Terms May Not Suffice for Microsoft Teams Communications
In Kim v. Cushman & Wakefield, a federal court held that search terms appropriate for email may not be sufficient for shorter, less formal communications on Microsoft Teams, requiring platform-specific search design.
- What happened
- In Kim v. Cushman & Wakefield, a federal court held that search terms appropriate for email may not be sufficient for shorter, less formal communications on Microsoft Teams, requiring platform-specific search design.
- Why it matters
- Organizations conducting eDiscovery collections must account for the distinct characteristics of collaboration platforms like Teams. Search terms designed for formal email communications may miss responsive material in shorter, informal messages, creating defensibility risk in production.
- Watch
- Decision dated April 24, 2026; discussed on EDRM blog June 22, 2026.
Sourcesedrm.net
Briefing item
Texas Business Court Holds ChatGPT Conversations Are Work Product but Orders Document Inventory
The Business Court of Texas held that a company executive's ChatGPT conversations about his litigation are protected work product, but ordered identification of every produced document fed into the tool, flagging an open protective order breach question.
- What happened
- The Business Court of Texas held that a company executive's ChatGPT conversations about his litigation are protected work product, but ordered identification of every produced document fed into the tool, flagging an open protective order breach question.
- Why it matters
- The decision establishes that AI tool use does not automatically waive work product protection, but practitioners must account for inventory obligations and potential protective order implications. Litigants using AI tools with case materials should implement tracking protocols before disputes arise.
- Watch
- Decision dated June 3, 2026; analysis published June 23, 2026. Protective order breach question remains unresolved.
Sourcesminerva26.com
Briefing item
Nextpoint and eDiscovery Today Examine Sanctions Risk from Incomplete TAR Productions
Brett Burney of Nextpoint and Doug Austin examined case law demonstrating that declaring a document production complete requires defensible validation, with flawed TAR workflows leading to motions to compel and Rule 37 sanctions.
- What happened
- Brett Burney of Nextpoint and Doug Austin examined case law demonstrating that declaring a document production complete requires defensible validation, with flawed TAR workflows leading to motions to compel and Rule 37 sanctions.
- Why it matters
- Counsel who certify production completeness without rigorous TAR validation face increasing sanctions risk. The discussion reinforces that quality control in technology-assisted review requires documented methodology, not merely reliance on automated classification outputs.
- Watch
- Published June 25, 2026. Video format; case law discussed not separately linked.
SourcesJD Supra
Educational only. Not legal advice. Review sources and matter-specific obligations before relying on any workflow decision.