Controlling and Coercive Behaviour: Widening the Net
Most of the commentary about the Singapore International Commercial Court (SICC) case of B2C2 Ltd v Quoine Pte Ltd  SGHC(I) 03 has focussed on the discussion of the nature of crypto-currencies/assets and whether they can be held on trust. However, the case also considers how the law relating to mistake could apply to contracts formed on an automated basis by computers using algorithmic software.
Broadly, there are two elements which must be proved in order to satisfy the Court that a contract should be voided on the basis of unilateral mistake. One party must prove that, without fault, it was mistaken as to one or more material terms in the contract and that the other party knew of, or should reasonably have known of, the mistake.
Quoine Pte Ltd (“Quoine”) operates an online platform for currency exchange where parties can trade cryptocurrency and fiat currency (the “Platform”). B2C2 Ltd (“B2C2”) was a user of the Platform. B2C2’s software included limits on the maximum and/or minimum price at which it was willing to buy/sell each particular currency. Quoine and B2C2 traded on an entirely automated basis through algorithmic trading software with no human involvement.
In April 2017, an oversight on Quoine’s part resulted in a number of trades taking place which involved the crypto-currency/asset called Ethereum being bought by the Platform at the lowest available price. After all lower priced Ethereum had been purchased, this resulted in B2C2 selling Ethereum for Bitcoin at a rate approximately 250 times higher than the going rate. The consideration in Bitcoin was automatically credited to B2C2’s account. The sales took place via the Platform, but the parties to the contracts were B2C2 and various other counterparties (“the Counterparties”).
The following morning, when the management of Quoine learned of the trades, the trades were reversed as the values deviated so greatly from the market rate.
In May 2017, B2C2 commenced legal proceedings against Quoine arguing that, inter alia the reversal of the trades was a breach of contract arising from a breach of the Platform’s terms and conditions.
Quoine defended the claim on the grounds that the contracts between B2C2 and the Counterparties were void on the basis of unilateral mistake.
The primary issue facing the SICC in respect of mistake in this case was, where contracts had been formed by computers, without human intervention, whose knowledge was to be assessed as to the mistake. In assessing this, the SICC asked itself the following questions:
“What mistakes have been made and to what extent are they fundamental? How does the Court assess knowledge or intention when the operation is carried out by computers acting as programmed? Whose knowledge is relevant? At what date is knowledge to be assessed?”
Whilst the SICC was only willing to develop the law in so far as it was necessitated by the particular facts, the approach it took to mistake in respect of automated contracts with no human intervention is interesting. The SICC concluded that the mistake must be made by the person on whose behalf the computer placed the order and the mistake must be in existence on or prior to the time of the trade.
The SICC found that there was a mistaken belief held by the Counterparties to the trades that those trades would be considered invalid as they were carried out at such an abnormal rate. Evidence from the Counterparties was that they believed the Platform would have sufficient safeguards to prevent trades being carried out in such circumstances. The SICC accepted the evidence of the Counterparties that they honestly held this mistaken belief and such a belief was fundamental to the contracts between the Counterparties and B2C2. The first element of unilateral mistake was therefore satisfied.
The SICC then considered whether B2C2 knew, or ought reasonably have known, of the Counterparties’ mistaken belief that they were protected against trades at an abnormal level, or that such trades would be treated as invalid.
In one of the most interesting parts of the judgment, the SICC determined that the knowledge to be assessed was the knowledge and intention of the programmer of the software (or the relevant part of the software).
The SICC heard evidence from the programmer of B2C2’s software. In order to prove that the programmer had actual knowledge that other users of the Platform held a belief that under no circumstances a trade would be completed at such deviant prices, he would need to hold the same belief at the time of programming the software. The programmer gave evidence that he included the upper and lower prices in B2C2’s software with the principal aim of protecting B2C2 from adverse trades. He did not exclude the possibility of trades being completed at those prices and he knew that the Platform was an automated system. On this basis, the SICC concluded that the programmer did not have actual knowledge of the mistaken belief. Further the SICC found that the programmer acted neither irrationally nor with impropriety and that a trader in his position would likewise not have held the necessary knowledge of the mistake.
The SICC therefore found that B2C2 had neither actual nor constructive knowledge of the mistake and Quoine’s defences of unilateral mistake at common law and/or in equity failed.
As pointed out above, the SICC made it clear that it was only willing to take the law as far as necessary in order to determine the case at hand. As such the precedent value of the case remains to be seen.
In this case it was relatively simple to identify the relevant software programmer for the purposes of investigating his knowledge and intention. However, this may become a far more difficult exercise where there is a team of programmers or where it is not clear which part or parts of the software caused an issue. It is foreseeable that in these situations the courts will need to hear considerable expert evidence on the issue.
An even greater issue likely to come to the fore in the not so distant future is the impact of Artificial Intelligence (AI) on the doctrine of mistake. The arguments put before the SICC may have been much more difficult to resolve if the higher and lower price limits in B2C2’s software were not pre-determined by the programmer but were instead decided by the computer itself. That being said, the SICC noted that the way in which ascertainment of knowledge of the mistake is determined in cases where computers have supplanted humans will be developed by the courts as and when disputes involving AI inevitably arise. The courts appear confident that they will be able to resolve such issues, as was stated by Lord Briggs in the case of Warner-Lambert Co Ltd v Generics (UK) Ltd  –
“The court is well versed in identifying the governing mind of a corporation, and, when the time arises, will no doubt be able to do the same for robots”.
The SICC has given us an indication of the approach the Courts may take in this case, and it will no doubt not be long before this is tested further.
Skip to content Home About Us Insights Services Contact Accessibility