What happened
Chauffeur drivers brought a wage-and-hour class action against J.E.T. Limousines & Transportation and other defendants in federal court in Arizona. The defendants proposed a technology assisted review protocol with a recall rate of 70-80%. The plaintiffs said that too many relevant documents would not be produced, and they asked for 90-95%. The defendants said that manual review after TAR, to raise recall, had delayed production. In an order dated 2 October and filed on 5 October 2026, Magistrate Judge John Z. Boyle declined to set a presumptive recall rate. He found insufficient evidence that a recall rate of 70-80% is deficient. He noted that such a rate is not out of the ordinary in federal litigation. He gave the defendants 75 days to finish the review. For two search terms, they had ten days to finish and to share validation metrics, including recall. The plaintiffs may raise recall again if they find that highly relevant documents were not produced. The court did not examine the review, which was still under way.
Sources
-
Court document, by authors with no tie to the system.
United States District Court for the District of Arizona (Magistrate Judge John Z. Boyle). McGhee v. J.E.T. Limousines & Transportation LLC, No. CV-24-03394-PHX-DJH, Order (Doc. 323), dated 2 October 2026 and filed 5 October 2026, pp. 3-4 and 8-9. CourtListener RECAP copy.
At the current juncture, there is insufficient evidence to conclude that a TAR recall rate of 70-80% is deficient, especially given that it is not out-of-the-ordinary to have such a TAR recall rate in federal litigation.
Details
AI system
- Workflow
-
The defendants proposed a TAR protocol with a recall rate of 70-80%. The defendants reviewed documents by hand after TAR to obtain a higher recall rate.
Follow-up
- Conditions for replication
-
The defendants give the plaintiffs the validation metrics for the review, including the recall rate. The plaintiffs can raise the recall rate again if highly relevant documents turn out not to have been produced.
Entry details
- Sector
-
Legal services / e-discovery.
- Purpose
-
A vendor that the order does not name built the TAR software to identify relevant documents for production.
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). Arizona court declines to raise a 70-80% recall target for TAR (AIBD-0990). Artificial Intelligence Benefits Database, edition of 1 November 2026. https://petervartanian.xyz/aibd/records/AIBD-0990.html (archived at https://doi.org/10.7910/DVN/AADSII, version 7.0)