What happened
In a business dispute between Astra and Musst, Musst gave disclosure under the disclosure pilot of the Business and Property Courts. Keyword searches cut 198,735 unique documents only to 175,723. Because of that volume, its then solicitors applied two types of technology assisted review, predictive coding and clustering analytics. They explained the steps over some three pages of the disclosure statement. Astra raised the use of technology assisted review in its applications. On 22 April 2020 Chief Master Marsh held that Musst should have notified Astra, and he called the omission a failure of process. He added that, in all probability, Astra could have had no proper basis to object once the estimated cost had been provided. He declined to revisit the point. The judgment does not state the cost or describe the software, and the court did not test the review.
Sources
-
Court document, by authors with no tie to the system.
High Court of Justice, Business and Property Courts of England and Wales, Business List (ChD) (Chief Master Marsh). Astra Asset Management UK Ltd & Anor v MUSST Investments LLB & Ors [2020] EWHC 1871 (Ch), judgment of 22 April 2020, paras 32 to 38. Find Case Law copy.
In all probability, had they been so informed there could have been no proper basis for objection once the estimated cost of using technology assisted review had been provided.
Details
AI system
- Workflow
-
Keyword searches cut 198,735 unique documents only to 175,723. Musst's then solicitors applied predictive coding and clustering analytics because of the volume. The disclosure statement explains the review steps over some three pages.
Follow-up
- Conditions for replication
-
A party that intends to use technology assisted review tells the other party and tries to agree on the approach. The other party receives the estimated cost of the review.
Entry details
- Sector
-
Legal services / e-discovery.
- Purpose
-
A vendor that the judgment does not name built the predictive coding and clustering analytics software to predict which documents are relevant and to group similar documents for review.
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). London court declines to revisit Musst's TAR of 175,723 documents (AIBD-0989). Artificial Intelligence Benefits Database, edition of 1 November 2026. https://petervartanian.xyz/aibd/records/AIBD-0989.html (archived at https://doi.org/10.7910/DVN/AADSII, version 7.0)