The High Court has approved the use of technology assisted review (TAR), including predictive coding, for electronic discovery in proceedings brought by Irish Bank Resolution Corporation Limited and others against Sean Quinn and others. The approval is subject to changes to the plaintiffs’ proposed protocol governing defendants’ participation.
Mr Justice Fullam ruled on 3 March 2015 that TAR using predictive coding can discharge a party’s discovery obligations under Order 31, rule 12 of the Rules of the Superior Courts, provided the process has sufficient transparency. The court stressed that no method is guaranteed to identify every relevant document.
The plaintiffs say the defendants conspired to wrongfully convert assets worth €455 million belonging to the former Anglo Irish Bank. That claim was background to the discovery application; the judgment did not decide the underlying allegation.
A March 2014 discovery order deferred a timetable while the plaintiffs assessed the volume of electronic material. Their initial keyword search identified 1.7 million potentially relevant documents. After duplicates and documents in other languages were removed, 680,809 documents remained suitable for predictive coding.
The plaintiffs argued that predictive coding would reduce the time and cost of review. Their evidence estimated a traditional review by 10 experienced reviewers would take nine months and cost €2 million, excluding supervision and technology. The defendants’ expert estimated a manual review by 10 reviewers could take 113 days and cost €220,000. The judge accepted that the proposed protocol would be more efficient in time and cost; the plaintiffs’ evidence was that fewer than 10 per cent of the remaining documents would need manual review.
The defendants objected that predictive coding might miss relevant material, questioned its suitability for the dataset and challenged how the process would be checked. The judgment describes safeguards in the protocol, including sampling documents below a relevance threshold, quality-control checks, and a process for challenging the plaintiffs’ coding of documents as not relevant. Defendants could also challenge the threshold in court.
The court found the plaintiffs had sought consent, shared the search terms and explanations of the method, and offered opportunities for input. It concluded that the defendants’ opposition did not prevent approval of a suitable protocol. Commercially sensitive, non-relevant documents would remain outside the disclosure check; the judge considered the proposed safeguards sufficient in the circumstances.
The protocol must provide for separate nominated counsel for the personal defendants and the Senat defendants. Each nominated counsel may discuss the PC Schedule with other defence counsel and attend the relevant meeting with one additional counsel from that defendant’s team. The ruling addresses the discovery method and protocol, not the merits of the underlying proceedings.