What happened
In a multidistrict antitrust case in the Western District of Pennsylvania, the defendants used search terms and technology-assisted review (TAR) to find documents for discovery. The plaintiffs moved to compel further TAR review. On 19 October 2022 E-Discovery Special Master James C. Francis IV recommended that Wanhua Chemical (America), Huntsman Corporation and Covestro resume their reviews. Each was to continue, presumptively, until the last two batches reviewed held no more than 10% responsive documents. He had understood that each company stopped while that share was materially greater than 10%. Huntsman's share was first given to him as 18%. On 21 October 2022 he reported that it was in fact 9.1%, a figure that the plaintiffs did not dispute. He found that Huntsman had already met his benchmark for a presumptively reasonable search and dropped the requirement for Huntsman. The district court adopted both reports on 7 November 2023. The supplemental report rests on a figure from a declaration filed in the case. It does not test the software, report recall or compare TAR with manual review.
Sources
-
Court document, by authors with no tie to the system.
Huntsman has already achieved the benchmark that I identified as representing a presumptively reasonable search, there is no longer a basis for
Details
AI system
- Workflow
-
Huntsman reviewed the documents from its TAR process in batches and then halted the review. The share of responsive documents in its last two batches was 9.1%.
Follow-up
- Conditions for replication
-
The last two batches reviewed contain no more than 10% responsive documents, the special master's benchmark for a presumptively reasonable search. The parties may still argue that a review should stop sooner or run longer, depending on how important the documents in those batches are.
Entry details
- Sector
-
Legal services / e-discovery.
- Purpose
-
A vendor that the report does not name built the TAR software to find the documents that answer discovery requests, which reviewers then check in batches.
31 more questions not yet checked
model, release, provider, role, configuration, inputs and outputs, autonomy, human oversight, AI contribution, beneficiaries, reach, distribution and access, duration, every measurement question, every money question, ownership and control, dependencies, cross-border connections, geopolitical implications, replication attempts, availability, institutional decision, grounds, action and result, success criterion, every evidence question, observed or intended, alternative assessments, related entries, other identifiers, category mappings.
Cite this record
Peter H. Vartanian. (2026). Huntsman's TAR review meets a special master's benchmark (AIBD-0994). Artificial Intelligence Benefits Database, edition of 1 November 2026. https://petervartanian.xyz/aibd/records/AIBD-0994.html (archived at https://doi.org/10.7910/DVN/AADSII, version 7.0)