China Metal Recycling (Holdings) Ltd (in Liquidation) v. Deloitte Touche Tohmatsu (A Firm)

HCA 1517/2016

[2022] HKCFI 2344

IN THE HIGH COURT OF THE

HONG KONG SPECIAL ADMINISTRATIVE REGION

COURT OF FIRST INSTANCE

ACTION NO 1517 OF 2016

________________

BETWEEN

________________

________________

D E C I S I O N

________________

1.The Liquidators of the China Metal Recycling (Holdings) Limited (“Company”) have brought a professional negligence action against, Deloitte Touche Tohmatsu (“Deloitte”), in respect of their audit of the Company’s financial statements for the 2009 to 2011 financial years.

2.Deloitte have applied for an order for a discovery protocol, which provides for limited discovery and the use of particular software to conduct that discovery. The material paragraph of the summons reads as follows:

“Leave be granted to the Defendant to give limited discovery as per the protocol set out in the Annex to this Summons and, in particular, that the Defendant be permitted to use a Technologically Assisted Review software technology, ‘Brainspace’ to reduce the time and costs of the discovery exercise.”

3.The purpose and thinking behind the application is succinctly explained in the introduction to Mr Yu’s skeleton argument[1].

(1)  The Summons is issued pursuant to the Court’s inherent jurisdiction and Order 24 rule 5 of the Rules of the High Court, Cap 4A (“RHC”), which gives the Court the power to make an order relating to the manner of discovery: see Hong Kong Civil Procedure 2022[2].

(2)  The Defendant proposes to use Technology Assisted Review (“TAR”) to facilitate its discovery process. The Plaintiff has agreed in principle to the Defendant’s use of TAR. TAR relies on a computer model to prioritise, but not replace, Clyde & Co’s manual review, where the data universe is over five million electronic documents. The value of TAR in litigation was recognised in case law in England, the United States and other common law jurisdictions. The Defendant has applied by summons dated 15 July 2022 to adduce the 3^rd affidavit of Paul Alexander Taylor (“Taylor 3^rd”) to update the Court on recent developments regarding the Defendant’s use of the TAR, and to confirm matters requested by the Plaintiff.

(3)  This is the first case where the appropriateness of the use of TAR will be considered by the Hong Kong Court.

4.RHC O24 r15A provides:

“For the purpose of managing the case in question and furthering any of the objectives specified in Order 1A, the Court may make any one or more of the following order—

(a) an order limiting the discovery of documents which the parties to the case would otherwise be required to make to each other under rule 1(1);

(b) an order directing that the discovery of documents required to be made under this Order to any party to the case shall, notwithstanding anything in this Order, be made in the manner specified in the order; and

(c) an order directing that documents which may be inspected under this Order shall, notwithstanding anything in rule 9 or 10, be inspected at a time or times specified in the order.”

5.The note to this Rule states:

“This new rule gives the court a wide power in the context of its duly to actively manage the case to make orders limiting discovery. Such orders may be made of the court’s own motion and without the necessity of an application by any party. Although the parties will be expected to agree their own discovery regime appropriate to the action (see the introductory comments at the beginning of the commentary on this Order) if they do not do so, or if the court considers a different regime more appropriate, the court may make an order accordingly.

The new rule also gives the court power to order the manner of discovery and the time for inspection.”

6.The reference to the Court’s duty to manage cases is a reference to RHC O1A, which contains a comprehensive, but not exhaustive, list of how this might be done. This expressly includes consideration being given to the use technology. I accept that the combination of RHC O1A and O25 r15A, allows a party to make an application to limit the scope of discovery. This is borne out by Sections 16.1 and 16.4 of the Final Report on Civil Justice Reform, which explain:

“478. [P]ractioners and the courts should instead be encouraged to use existing RHC powers to fashion a discovery regime suitable to the needs of the particular case—preferably by agreement, but otherwise by order. Instead of cases being routinely allowed to proceed to full automatic Peruvian Guano discovery, it ought to become standard practice to consider whether any economising modifications should be made to the scope and manner of meeting the parties’ discovery obligations. A practice direction in suitable terms should be issued and a question prompting such economies should be included in the timetabling questionnaire.[3]

…

500. As indicated in the discussion of Proposals 25 and 26 above, there was a broad consensus that the excesses of discovery ought to be tackled by appropriate case management by the courts. The balance of opinion favoured taking full Peruvian Guano discovery as the starting point, to be narrowed by appropriate case management, rather than (as Proposal 29 suggests) starting from ‘standard discovery’ and deciding when to widen its scope. No further discussion is called for in the light of the foregoing Recommendations.

7.It is not correct to say that the Court has not previously had to consider the appropriateness of the use of TAR. TAR was used in the litigation between Elliott International LP v Bank of East Asia Ltd (No 1)[4]. My experiences of dealing with discovery in that matter inform my views on the present application. TAR was used by the Bank to search a massive quantity of data that might possibly contain relevant documents. The Bank did not seek the Court’s approval to the use of TAR and Elliott did not object to its use or suggest that it should not have been used without the Court’s prior approval. It seems to me that it is not necessary for the Court to approve the use of TAR. Parties to proceedings involving discovery are under an obligation to search for relevant documents (in practice these days data stored electronically rather than pieces of paper) and how this is to be done will to some extent depend on the nature of the data and the resources available to the parties and their lawyers. The Court does not direct how a party is to conduct traditional discovery so as to ensure compliance with a party’s discovery obligation. In my view there is no particular reason why it should be necessary for the Court to approve the use of TAR or any other technology, which a party’s lawyers believe will facilitate discovery, although it is sensible for the parties to agree in advance how TAR is to be used. This is what has happened in the present case. Over time the parties have agreed nearly all of a protocol, that process culminating in the Liquidators’ counsel skeleton, which leaves the Court with little to decide.

8.I have some reservations about the Court involving itself at the outset of discovery in the determination of the detailed methodology for the use of TAR or similar technology. There are two main reasons for this. First, the Court does not have the Parties’ familiarity with the minutiae of the case and its preparation, which is often what needs to be understood in order to assess the problems that discovery poses and how they are best addressed. Secondly, the Court generally does not have practitioners’ understanding of the technology available and the ways in which it can be used. It is will often be more practical for the parties to agree as much of the methodology for undertaking discovery as they can, undertake discovery and only involve the Court if there are disputes over the adequacy of discovery. The method of carrying out discovery is a means to an end. What method is used is not important if the end is successfully achieved.

9.In deciding to make this application Deloitte have been influenced by two English decisions, Pyrrho Investments Ltd v MWB Property Ltd[5] and Triumph Controls UK Ltd v Primus International Holding Co[6]. The import of the two decisions is that it is prudent for a party proposing to use TAR to obtain the Court’s agreement first. In Triumph Controls this view seems to have been driven by the fact that a sampling method was to be used, which necessarily would not involve an exhaustive search and assessment of documents.

10.Deloitte’s evidence explains how the TAR software it proposes to use, Brainscan, will identify documents (which necessarily are in digital form), which may be relevant through use of a machine learning algorithm, which develops criteria which results in a score (between 0 and 1) being allocated to a document. A score of 1 indicates a document falls into the most relevant category; 0 is the least relevant. The aim is not to replace a manual review, but to allow prioritisation of the checking by the legal team of documents, which Brainscan identifies as of most relevance. It may be that if sample manual checking of documents identified by Brainscan as likely to be of little or no relevance demonstrates that Brainscan is highly accurate in assessment of relevance, further documents identified as of little relevance will not be manually checked. Those documents identified by Brainscan to be of relevance will then be scrutinised by Deloitte’s lawyers.

11.It seems to me clear that there is nothing objectionable in documents being identified for relevance during discovery using TAR. There is nothing in the Rules of High Court, which suggests otherwise and, as I have noted, RCH O1A envisages the use of technology where it is beneficial to do so. Practice Direction SL1.2 [4(3)] and [24] requires it in the Commercial List in the case of discovery of what it refers to as Electronic Documents, which is data stored in electronic form, i.e., digital form.

12.Unsurprisingly, a similar approach is taken by other courts. Judge Andrew Peck in Da Silva Moore v Publicis Groupe[7] observes that the computer assisted review “should be used where appropriate”, noting that it is not a case of “machine replacing humans”. The Judge observed from practitioner and academic studies that “computerised searches are at least as accurate, if not more so, than manual review” (see 190). It was held that the use of computer assisted review was appropriate, given “(1) the parties’ agreement, (2) the vast amount of ESI to be reviewed (over three million documents), (3) the superiority of computer-assisted review to the available alternatives (i.e., linear manual review or keyword searches), (4) the need for cost effectiveness and proportionality” (see 192).

13.In a subsequent case, Rio Tinto v Vale[8], Judge Peck observed that “[i]n the three years since Da Silva Moore, the case law has developed to the point that it is now black letter law that where the producing party wants to utilize TAR for document review, courts will permit it” (at 127). It was noted, in the context of whether an opponent is entitled to access to the “seed set” of documents used to the train the model, that the “seed set” is “much less significant” if the TAR Model uses “continuous active learning” (at 128). In both Australia and New Zealand[9] the use of TAR in cases involving large quantities of documents is approved and encouraged.

14.In my view the issue is not whether TAR can be used. Clearly in 2022 the use of analytic tools of this sort is to be expected in the conduct of litigation whether to research the law or to order and analyse data. I can see no reason why the default position should be assumed to be that a person with some legal qualification personally checks data on a screen. Increasingly all data will be stored and accessed through electronic means. There will be no paper to look at unless a lawyer chooses to print out data. The issue is likely to concern the extent to which the scope of discovery is reduced to something less than wide Peruvian Guano type discovery and the interface between this issue and the use of TAR.

15.The present case involves a substantial amount of data. As the audit took place about 10 years ago there are 300 hard copy bundles of documents and over five million electronic documents. Clearly discovery is a major exercise and it is sensible for technology to be used in order to allow it to be conducted as efficiently and cost effectively as possible. I, therefore, do not agree with the Liquidators’ submissions (although, I note they do not oppose the use of TAR) that Deloitte bears the burden to satisfy the Court that its protocol should be sanctioned, because in my view Deloitte does not require the Court’s sanction to use it. However, as the application has been made and the protocol very largely agreed, I will make directions subject to one minor amendments which is the time within which discovery is to be completed by Deloitte.

16.Deloitte seeks an extension of time of 150 days. The Liquidators are prepared to agree 60 days. Deloitte says that discovery is immense and the additional time is simply a consequence of this and to some degree the difficulty in agreeing a protocol to use TAR. As the Liquidators recognise an extension of time is inevitable. Discovery was ordered to be completed by 17 January 2022, a date proposed by Deloitte on 2 June 2021. Deloitte have had a de facto extension of six months. It now seeks a further five months. It seems to me this is excessive. There can be little science in assessing what is reasonable in the circumstances. Deloitte needs to deal with discovery with urgency. I will make a final order that discovery takes place on or before 4:30pm 31 October 2022.

Costs

17.Although the present application was issued for sensible reasons, it does involve an application for an extension of time and an order necessary to facilitate Deloitte’s discovery. I will order that the costs are the Liquidators’ (Plaintiff) costs in the cause.

Mr Justin Ho, instructed by Holman Fenwick Willan, for the plaintiff

Mr Jason Yu, instructed by Clyde & Co, for the defendant