Lead story
New Whitepaper: Evaluating Methods for Review in eDiscovery Today
Released by eDiscovery Today for its 6-year anniversary, the whitepaper is based on a September 2025 webinar. It covers different approaches to document review (TAR, CAL, generative AI, manual), current trends, and how to evaluate them. Addresses whether keyword search with manual review is still effective, obligations under Sedona Principle 6, and how to measure review effectiveness to ensure defensible outcomes.
- What happened
- Released by eDiscovery Today for its 6-year anniversary, the whitepaper is based on a September 2025 webinar. It covers different approaches to document review (TAR, CAL, generative AI, manual), current trends, and how to evaluate them. Addresses whether keyword search with manual review is still effective, obligations under Sedona Principle 6, and how to measure review effectiveness to ensure defensible outcomes.
- Why it matters
- As review methods multiply, practitioners need frameworks to validate their chosen approach. This provides practical guidance on demonstrating reasonableness to courts and avoiding sanctions.
- Watch
- Review methodology evolves quickly; ensure any framework aligns with specific jurisdiction expectations and case requirements.
SourceseDiscovery Today
Briefing item
Case Where Plaintiff Tried to Use ChatGPT in Deposition is Dismissed
*Jones v. Delta Air Lines, Inc.*, No. 2:24-cv-11224 (E.D. Mich. Apr. 22, 2026). A pro se plaintiff used ChatGPT to assist in answering deposition questions. Judge Susan K. DeClercq issued a clear directive prohibiting AI assistance during testimony. The case was dismissed with prejudice under Rules 37(b)(2) and 41(b) due to repeated failure to comply with discovery obligations, though the AI issue was notable but not the primary basis for dismissal.
- What happened
- *Jones v. Delta Air Lines, Inc.*, No. 2:24-cv-11224 (E.D. Mich. Apr. 22, 2026). A pro se plaintiff used ChatGPT to assist in answering deposition questions. Judge Susan K. DeClercq issued a clear directive prohibiting AI assistance during testimony. The case was dismissed with prejudice under Rules 37(b)(2) and 41(b) due to repeated failure to comply with discovery obligations, though the AI issue was notable but not the primary basis for dismissal.
- Why it matters
- This is one of the first documented cases of a party attempting to use GenAI during a deposition. The judge's explicit prohibition establishes early precedent for AI boundaries in testimony.
- Watch
- Dismissal was based on broader discovery failures; jurisdictional variation expected as more cases emerge.
SourceseDiscovery Today
Briefing item
Agentic AI Liability: Managing Accountability in Autonomous Legal Workflows
Scott Cohen (Lineal EVP) addresses the shift from assistive AI to 'agentic AI'—systems that act autonomously across workflows. Key issues: supervision becomes harder when work is ongoing and system-driven, agents can operate beyond intended boundaries, and malpractice carriers are introducing exclusions for higher-autonomy use cases. ABA Model Rules 1.1, 5.1, and 5.3 still apply regardless of AI autonomy.
- What happened
- Scott Cohen (Lineal EVP) addresses the shift from assistive AI to 'agentic AI'—systems that act autonomously across workflows. Key issues: supervision becomes harder when work is ongoing and system-driven, agents can operate beyond intended boundaries, and malpractice carriers are introducing exclusions for higher-autonomy use cases. ABA Model Rules 1.1, 5.1, and 5.3 still apply regardless of AI autonomy.
- Why it matters
- As eDiscovery platforms integrate autonomous capabilities (auto-coding, smart batching), firms need governance frameworks with validation checkpoints, audit logs, and escalation triggers to manage liability and insurance compliance.
- Watch
- Insurance exclusions for agentic AI may impact firm risk profiles; ensure coverage reviews align with implementation plans.
Sourcesaceds.org
Briefing item
AI Work Product and Protective Orders: Morgan v. V2X, Inc.
Addresses whether AI-generated work product is discoverable under protective orders, highlighting the tension between AI efficiency and privilege/work product protections.
- What happened
- Addresses whether AI-generated work product is discoverable under protective orders, highlighting the tension between AI efficiency and privilege/work product protections.
- Why it matters
- Early guidance on how courts may treat AI-assisted analysis under protective orders—a growing concern as TAR and AI review become standard in litigation.
- Watch
- This appears to be a developing issue without settled precedent; monitor for circuit splits.
Sourcesaceds.org
Briefing item
Relativity to Establish Singapore Entity, Expanding APAC Footprint
Relativity plans to launch in Singapore while accelerating AI-driven legal data intelligence adoption across the APAC region, highlighting innovation at RelFest Sydney.
- What happened
- Relativity plans to launch in Singapore while accelerating AI-driven legal data intelligence adoption across the APAC region, highlighting innovation at RelFest Sydney.
- Why it matters
- Growing APAC presence indicates maturing global eDiscovery market and importance of regional AI capabilities with localized regulatory compliance.
- Watch
- Verify specific timeline and services that will be localized for APAC markets.
SourceseDiscovery Todayedrm.net
Educational only. Not legal advice. Review sources and matter-specific obligations before relying on any workflow decision.