I. Introduction
Digital assetsFootnote 1 have been gaining traction in recent years and various common law jurisdictions now recognise them as objects of property rights.Footnote 2 At the moment, however, there is no legal regime that deals with interferences with digital assets: the “chattel torts” are only applicable to assets that are amenable to possession and thus only apply to tangible assets, and other causes of action only offer piecemeal protection for digital assets. Where someone causes a digital asset to be frozenFootnote 3 or burned,Footnote 4 there is very little certainty as to what remedies would be available to a claimant. Considerable uncertainty also exists in many situations where someone has been denied access to a digital asset.Footnote 5 This uncertainty is unacceptable, as it encourages people to interfere with digital assets, drives up the cost of litigation and creates the risk of defendants being able to strong-arm individuals into settling for a low sum.
On the one hand, it has been argued by various academics that the tort of conversion should be extended to cover digital assets.Footnote 6 Similarly, the Law Commission in their Consultation Paper noted that there is a “good argument for extending the tort of conversion”Footnote 7 to digital assets. This is on the basis that digital and physical assets are similar enoughFootnote 8 such that it would be arbitrary not to subject them to the same interference regime. Without such an interference regime being applicable to digital assets, owners and holders of digital assets would be insufficiently protected. Also, although the arguments have focused specifically on the tort of conversion, it is not only conversion that needs to be extended to cover digital assets if the full spectrum of “equivalent”Footnote 9 interferencesFootnote 10 is to be covered. The other “property torts” or “chattel torts” of trespass and reversionary injury would also need to be extended to cover digital assets.Footnote 11
On the other hand, this extension has not been supported in the most recent literature. For example, the Law Commission in their Final Report has taken the view that conversion in its current form should not be applied to digital assets, revising their stance from that taken in the Consultation Paper.Footnote 12 They reached this view on the basis that physical and digital assets behave very differently and that there are potential issues with the strict liability nature of conversion.Footnote 13 Similarly, the Dubai International Financial Centre (DIFC) considered but rejected the proposal to extend the chattel torts to digital assets, opting instead in favour of a new bespoke regime that deals with digital asset interferences.Footnote 14
This article agrees with the view in the Law Commission Final Report as well as the DIFC Consultation Paper that the chattel torts should not be extended to digital assets.Footnote 15 It is suggested that this is for two reasons. First (and most fundamentally), physical and digital assets are very different in their nature, behaviour and respective environments, meaning that the concepts and thresholds used in the chattel tort context cannot be usefully applied in the digital asset context. This, coupled with the fact that digital assets are an asset class that judges tend to be substantially less familiar with,Footnote 16 creates uncertainty and a very substantial risk of producing the wrong normative result.Footnote 17 Second, the chattel tort rules themselves are unsatisfactory, needlessly complex and create problems for innocent defendants, and so applying such rules across to digital assets will mean that these negative characteristics will be replicated in the digital asset interference context.
This conclusion would be useful to a legislature, court or law reform body deciding how best to protect against interferences with digital assets. In order to know which means of protection would be best, one needs to know the problems with the chattel torts (and with applying them to digital assets), in order to compare the pros and cons of extending the chattel torts to digital assets against the pros and cons of any alternative means of protection proposed (e.g. a regime that is similar to that contained in Articles 14 and 15 of the DIFC’s Digital Asset Law (DAL) 2024).Footnote 18 The purpose of this article is to contribute to such an exercise by (1) showing how extending the chattel torts to digital assets gives rise to many problems and is undesirable, and thereby (2) providing a significant reference point against which other means of protection can be compared.Footnote 19
I will first outline the limitations of the chattel torts and how these limitations give rise to a gap in protection in respect of digital assets (Section II) and then describe the argument in favour of extending the chattel torts to digital assets as well as briefly outline the two substantive arguments against doing so (Section III). I will then discuss the structure and general features of an interference with a physical asset (Section IV). Next, I will discuss in more detail the two substantive arguments against extending the chattel torts to digital assets (Sections V and VI). Finally, I will discuss why the existence of “digital trespass and conversion” cases in other jurisdictions does not negatively impact the strength of the two substantive arguments against extending the chattel torts to digital assets (Section VII).
II. Possession, Intangibles and the Gap in Protection
At present, the chattel torts (conversion, trespass and reversionary injury) only apply to assets that can be possessed, and possession at present just applies to tangibles.Footnote 20 This means that the chattel torts do not apply to digital assets.
This carries various implications. The primary implication is that various core cases of the claimant’s use of his digital assets being impaired (e.g. the defendant freezing or burning the claimant’s digital asset) generally do not give rise to a remedy under English law as it currently stands,Footnote 21 whereas the equivalentFootnote 22 impairment in respect of a physical asset gives rise to a claim under the chattel torts.Footnote 23 Without extending the chattel torts to digital assets, we would need to rely on the economic torts, unjust enrichment, the intellectual property torts and the information torts; however, these causes of action offer limited protection to claimants in such a situation.
If a defendant freezes or burns a digital asset, the economic torts may not be of assistance. First, inducing breach of contract requires a prior contract, which may not exist. Second, the “causing loss by unlawful means” tort requires a prior civil wrong committed by the defendant against a third party that affects the liberty of such a third party,Footnote 24 which may not exist especially if the defendant is interacting directly with the blockchain. Third, deceit requires the defendant to make a representation and for the claimant to rely on it,Footnote 25 which would not apply in cases where the interference does not require the cooperation of the claimant and in cases where the defendant makes no representation to the claimant. Fourth, other economic torts such as conspiracy or intimidation would clearly be inapplicable in situations where the defendant acts aloneFootnote 26 or if there are no threats involved.Footnote 27
Unjust enrichment is also of limited assistance, because neither freezing nor burning involves enrichment to the defendant. In any event, there are issues with establishing the unjust factor.Footnote 28 Likewise, the intellectual property torts (e.g. copyright infringement) and the information torts (e.g. breach of confidence or misuse of private information)Footnote 29 may not assist in cases of freezing and burning because there is usually no intellectual property right that is being infringed,Footnote 30 and there is usually no use or disclosure of confidential or private information.Footnote 31
As such, one may wonder whether the best solution to fill this gap would be to extend the chattel torts such that they apply in the digital asset context.
III. The Argument for Extending the Chattel Torts and the Two Arguments Against
The argument that conversion should be extended to intangible assets has been commonly made.Footnote 32 In respect of digital assets specifically, this argument has chiefly taken the form of the “anomaly” argument. Specifically, since both physical and digital assets can be stolen, transferred,Footnote 33 and are objects independent of the legal system, it would be anomalous to treat them differently. If misappropriating an iPhone (in a way that the claimant can no longer access it) constitutes conversion, so should misappropriating Bitcoin (in the form of an unauthorised transfer to a different blockchain address, such that the claimant can no longer access it). The similarity between physical and digital assets has been noted by the Law Commission as well as various academic commentatorsFootnote 34 who make the argument that they should be treated in like manner for the purpose of conversion (and presumably also the other chattel torts).Footnote 35
Sarah Green and Ferdisha Snagg, for example, argue that there has been an over-emphasis on tangibility, noting that tangibility is merely a proxy for the distinction between “abstract” and “concrete” things: “tangibility historically described those things that were concrete, but it does not follow that it had any determinative influence on that categorisation.”Footnote 36 The normatively significant distinction for the law’s purposes is that between abstract and concrete things. Abstract things do not have an existence independent of the legal system and relationships between individuals (e.g. debts), but concrete things do have such an existence (e.g. tables, chairs and cryptosecurities).Footnote 37 It is this distinction (as opposed to the distinction between tangibility and intangibility) that should be determinative in deciding whether the chattel tort regime applies to a particular type of asset.
This is because concrete things, unlike abstract things, are “excludable and exhaustible” in the sense that they can be lost and stolen (exhaustible)Footnote 38 and capable of exclusive control (excludable) regardless of whether a legal system exists and regardless of whether anyone claims rights in relation to them.Footnote 39 If an asset is excludable and exhaustible, “it can be possessed in a legal sense”.Footnote 40 This therefore allows one to group digital assets together with chattels insofar as both types of assets are concrete things,Footnote 41 thus bringing in the protections of the chattel torts.Footnote 42 As such, the chattel tort regime should apply to digital assets (which are concrete things).
This argument for applying conversion (and by extension the chattel torts)Footnote 43 to digital assets may be further bolstered by the fact that the Law Commission has proposed the extension of possession to electronic trade documents, meaning that electronic trade documents can be converted.Footnote 44 If “possession” applies to electronic trade documents, it would (on this argument) be arbitrary not to apply the concept of possession to digital assets generally.
However, even if some digital assets satisfy this statutory definition of possession, it does not automatically follow that the same concept or term (possession) applies (1) generally as a matter of common law and (2) to all digital assets. In this statutory context, “possession” is used for a particular purpose and the legislation gives effect to a specific policy: to eliminate the differential treatment between physical trade documents and digital/electronic trade documents, given that they serve the same purpose in commerce.Footnote 45 In contrast, applying the same concept (possession) to digital assets generally carries much wider implications, as many doctrines depend on the applicability of the concept (e.g. bailment, possessory security, delivery and the chattel torts) and the normative balance is different in the context of digital assets. The stakes are also much higher, since extending possession to digital assets as a matter of general common law creates the risk of distorting the law.Footnote 46
Nonetheless, other jurisdictions apply the chattel torts to intangibles: for example, in the US, there are cases applying conversion and trespass to digital assets.Footnote 47 There is also case law in the US, Canada and New Zealand applying conversion in the context of non-crypto “digital assets”Footnote 48 such as digital files and domain names.Footnote 49
However, it is suggested that the arguments for extending the chattel torts to digital assets are insufficiently focused on the big picture. Specifically, they assume that physical and digital assets share enough similarities that they can be treated in the same way for the purpose of the interference torts. This assumption breaks down when one considers the fundamental differences between the two types of assets, which have been noted by the Law Commission in their Final Report, as well as the DIFC in their DAL Consultation Paper.
Both the Law CommissionFootnote 50 and the DIFC believe that conversion and/or the chattel torts should not be applied to digital assets, because the two types of assets are so fundamentally different. The Law Commission notes that chattels and digital assets “behave in different ways”,Footnote 51 such that applying conversion in the digital asset context would not be desirable.Footnote 52 Similarly, in the DAL Consultation Paper, the DIFC notes that physical and digital assets are “very different in nature and surrounding environments”, meaning that it would be difficult to avoid “creating unacceptable uncertainty or substantially increasing the risk of incorrect decisions” if the chattel torts were to be applied to digital assets. This corresponds with the first substantive argument of this article (explored in Section V).Footnote 53
The DIFC gives another reason why the chattel torts should not be extended to digital assets. It notes that the existing chattel tort regime is “unsatisfactory and needlessly complex”,Footnote 54 such that if it were to be extended to digital assets, the same “undesirable features [would be] replicated in the [d]igital [a]sset context”.Footnote 55 This corresponds with the second substantive argument of this article (explored in Section VI).Footnote 56
In order to contextualise the two substantive arguments of this article, it will be useful to set out the general structure and elements of an interference with a physical asset.
IV. The Interference Regime for Physical Assets
There is a general structure to every case of interference with a physical asset. First, there is an action taken by the defendant. Second, the action impacts the claimant’s asset or his use of the asset (impact). Third, the link between the action and the impact is proximate enough (proximity/causation).Footnote 57 Fourth, the defendant also has a mental state when he performs the relevant action. Finally, we need to look at what constitutes a defence to the cause of action, even if the first four elements are satisfied.
In order to ascertain whether the chattel tort regime can be transposed into the digital asset context, we need to analyse the five elements in respect of physical asset interferences and determine the consequences of applying the same threshold to digital assets.
The five elements will be explored in turn.
A. Elements 1 and 2: Action of the Defendant and Impact on the Claimant’s Asset or Use of His Asset
The first two elements will be discussed together because in the context of the chattel torts element 1 (the action of the defendant) forms part of the definition of element 2 (the impact on the claimant or his asset).
The duty on the defendant consists of a duty not to (deliberately) physically interfere with the claimant’s asset.Footnote 58 In the context of elements 1 and 2, the requirement that the defendant must not physically interfere with the claimant’s asset can be broken down into a few subduties. First, he must not take any deliberate positive action that physically damages the claimant’s chattel. Second, he must not make any deliberate physical contact with the asset (which may happen through using his own body or through other means, such as through an object). Third, he must not completely impair the claimant’s use of his asset (irrespective of whether there has been any physical contact or damage).Footnote 59 Fourth, he must not enter a transaction that deprives the claimant of his title to the asset.Footnote 60
These four duties are covered by the three torts of conversion, trespass and reversionary injury.
Conversion covers the most severe types of interference. Specifically, it requires a deliberate action by the defendant that either (1) physically damages the chattel (or involves physically touching the chattel) in a way that totally or severely excludes the claimant from use of the chattelFootnote 61 (even for a temporary period)Footnote 62 or (2) directly excludes the claimant from using the chattel (for any period of time) despite a lack of physical contact or damage. Examples of (1) would include taking, selling,Footnote 63 or destroying the asset, or detaining the asset with an intention to assert title,Footnote 64 or transforming the asset in a way that it loses its essential identity.Footnote 65 As for (2), this includes a sale that deprives the claimant of his title,Footnote 66 and this results in a “total exclusion of use” in the sense that the claimant’s use of the asset would involve the incurring of a liability to the new owner (given that he no longer has title).Footnote 67
Indeed, partial impairments of use do not constitute conversion in the absence of physical contact, as demonstrated by Club Cruise Entertainment and Travelling Services Europe BV v Department for Transport; The Van Gogh.Footnote 68 In Club Cruise, an official served an administrative detention notice on the claimant shipowner, mandating it to stay in port. The detention notice turned out to be invalid and the claimant sued in conversion. The court held that there was no conversion, since there was no physical contact or restraint and the claimant still had possession of the ship.
Trespass requires intentional action by the defendant that physically touches or damages the chattel (and covers lesser interferences that are not serious enough to amount to conversion).Footnote 69
Nonetheless, a person can only sue in conversion if he had actual possession or a right to immediate possession of the chattel at the time of the interference and can only sue in trespass if he had possession of the chattel at the time of the interference.Footnote 70 As such, many people with a reversionary interest in a chattel (e.g. a pledgor or a term bailor) would not be able to sue in conversion or trespass.
This is where the tort of reversionary injury becomes relevant. It covers any act that would constitute conversion, trespass (or negligence), but also requires actual damage,Footnote 71 and is only available to someone who has a reversionary interest in a chattel. A person with a reversionary interest will be able to sue in reversionary injury if there has been damage to his interest, as opposed to damage to merely the pledgee or bailee’s interest.Footnote 72
It is worth noting that not every impairment or change in form amounts to an interference. For example, Simon Douglas gives the example of a person who buys up all the local supplies of petrol.Footnote 73 This would lead to people’s use of their cars being impaired, as the cars would not have fuel anymore and so people would not be able to drive their cars. However, this impairment is not interference for the purpose of the chattel torts, because, if such an action attracted liability in the chattel torts, this would unduly limit the liberty of defendants.Footnote 74
Overall, physical contact or physical damage is a requirement for trespass and for conversions that do not amount to a total impairment of use. It is also a requirement for the equivalent reversionary injury claim,Footnote 75 provided there is actual damage.Footnote 76 If there is no physical contact with or physical damage to the chattel, the conduct requirement is increased (i.e. there must be a total impairment of use before there can be an interference).Footnote 77 Nonetheless, in most physical asset interference cases, physical contact or damage usually exists.
B. Element 3: Proximity/Causation
There is a directness requirement for trespass,Footnote 78 meaning that there is a causation constraint that limits the defendant’s liability. A defendant does not commit trespass if he lays a trap for a physical object to fall into.Footnote 79 As such, situations that potentially engage the “intervening acts of causation” debate are outside the scope of the tort.
In respect of conversion, the types of actions that constitute the tort all involve a very direct causal chain (e.g. taking or destruction of an asset).Footnote 80 There is no intervening act between the defendant’s action and the impact on the claimant (or his asset), because the acts that involve conversion involve one of two patterns, both of which do not involve any act in between the defendant’s action and the impact on the claimant (or much time in between). First, there are cases involving direct contact with the chattel (where the defendant’s action and the impact on the asset happen at the same time (or almost at the same time)).Footnote 81 Second, there are cases involving a total exclusion of the claimant’s use of the chattel without any act in between the defendant’s action and the claimant’s exclusion from use (such as Burroughes v Bayne).Footnote 82 As such, the “intervening acts of causation” debate is not relevant in the conversion context.Footnote 83
These “causation” principles also apply in respect of reversionary injury, as the tort covers the same ground as conversion and trespassFootnote 84 insofar as deliberate interferences are concerned,Footnote 85 provided that a reversionary interest is damaged.Footnote 86
C. Element 4: Mental State
The chattel torts impose strict liability.Footnote 87 This has been criticised on the basis that it is overly harsh on defendants and does not provide “fair warning”,Footnote 88 but one could also justify it on the basis that elements 2 and 3 are narrowly constrained. Specifically, because physical contact or physical damage is a requirement where there is no total impairment of use, this provides some degree of fair warning. The boundaries of the physical thing provide a crucial limit to the potential scope of liability,Footnote 89 which reduces (or eliminates) the need for liability to be constrained by way of a mental element.
D. Defences
There are various specific statutory defencesFootnote 90 and in terms of general common law the main defence is consent.Footnote 91 If the claimant expressly or impliedly consents to the interference, the defendant has a defence.Footnote 92
V. Argument 1: Difficulty in Applying Chattel Tort Elements and Normative Balance to Digital Assets
The most fundamental reason why the chattel torts should not be applied to digital assets is that there are many significant differences between the nature, behaviour and environment of physical and digital assets. As a result, existing concepts that are used to resolve disputes in the physical asset context are not adequate to resolve disputes in the digital asset context.
There are various differences between physical and digital assets that are worthy of note. First, physical assets have a distinct molecular boundary that defines the space that they occupy, whereas digital assets have no distinct molecular boundary. Second, the blockchain environment is an “opt-in” environment that one has the option not to join, whereas the physical environment is something that we are part of no matter what. Third, the blockchain environment is “composable” in the sense that coders can define the features of the “blockchain world” they create to a much greater degree than a person in the physical world can define the features of the assets they create.Footnote 93 Fourth, a digital asset can only be accessed through a digital device, whereas a physical asset can be accessed by making contact with the (physical) space in which it is contained. Fifth, we have a much better idea of how physical assets work and behave, given that we interact with them on a daily basis and they occupy a molecular space that our senses are attuned to, meaning that we can spot dangers arising out of them by using our senses (primarily vision and touch). This stands in contrast with digital assets where we are unable to spot similar dangers, given that the code-governed environment is unfamiliar to most people and many cases of impairment may occur outside our knowledge or foresight.
Because of these differences, concepts that are adequate to resolve disputes in the physical asset context are not adequate to resolve disputes in the digital asset context. They do not yield determinate resultsFootnote 94 in the digital asset context: for example, in the case of concepts such as physical interference or physical contact, the nature and environment of digital assets is such that no direct analogy with physical assets can be drawn.Footnote 95
Yet, a judge still needs to make a decision on a given set of facts and so he or she will need to try and find the equivalent of physical interference or physical contact. However, there will not be a precise equivalent and whichever equivalent is applied will be (at best) a rough approximation. We also do not know which approximation the judge will apply, given that there are many possible options (explored below).Footnote 96
This creates an unacceptable amount of uncertainty for parties, given that they will find it extremely difficult to predict their legal positions (since each proxy or approximation generates a substantially different scope of liability)Footnote 97 and parties will need to litigateFootnote 98 in order to find out their legal positions.Footnote 99 This creates a substantial risk of a chilling effect on users of the blockchainFootnote 100 (including centralised exchanges and operators of blockchains) as they may fear liability under the chattel torts if they take certain digital actions (and, in particular, actions on the blockchain).Footnote 101
Apart from the problem of uncertainty, there is a significant risk that judges will produce the wrong normative threshold by picking an inaccurate equivalent. Since digital assets are technically complex and thus difficult to understand, judges may be misled into using an inaccurate proxy that produces an overly wide or narrow scope of liability. Alternatively, judges may pick a proxy that can no longer be seen as an equivalent to the corresponding chattel tort requirement, which effectively modifies the threshold in respect of the chattel torts while paying lip service to the requirement in question.
Judges may also produce the wrong normative threshold through directly applying the physical asset threshold to digital assets. This is because the differences in physical and digital assets (and the different policies at work) justify different requirements and applying the same requirement produces undesirable results in the digital asset context. An example that will be explored below relates to the role of digital asset “kill switches”Footnote 102 in the context of the consent defence.Footnote 103
A. No Proxy for Physical Interference
First, in the digital asset context, there is fundamentally no equivalent of, or proxy for, the physical contact/damage requirement for trespasses and for conversions that do not lead to a total impairment of use.Footnote 104
Physical assets have physical boundaries (i.e. molecular boundaries) and the physical contact/damage requirement provides a very important limit to the scope of liability, especially given that the chattel torts attract strict liability. It is relatively easy to identify whether there has been physical contact (touching of molecules that constitute the chattel) or physical damage (a change in the molecular structure of the chattel that renders the chattel less useful or valuable).Footnote 105 This also allows a clear distinction to be drawn between a physical interference and an impairment of use: there can be one without the other.
There is no equivalent of physical contact or physical damage to (or physical interference with) a digital asset. A digital asset is ideational and common ways in which destruction or denial or impairment of access occur include where the defendant causes (1) the freezing of the asset, (2) a “denial of service” attack, (3) a transfer of the asset to another address (whether it is a smart contract or wallet address and whether it has a private key or not)Footnote 106 or (4) destruction of the asset through burning. There is no equivalent of a physical boundary that people can walk into and interact with, because of fundamental differences in the nature of physical and digital assets. These differences mean that, unlike physical assets where one can cleanly distinguish impairment of use and physical interference, one cannot cleanly distinguish impairment of use of a digital asset from digital interference with the asset.
If one applies the physical damage/contact requirementFootnote 107 literally to digital assets, this will result in the claimant having no remedy in situations that involve freezing his digital asset in a way that partially impairs the use of his digital asset. This is because there is no physical damage/contact, which means there is no liability in trespass or conversion.
Nonetheless, are there any proxies that can serve as adequate substitutes for such a requirement? One possibility would be to limit actionable interferences to “on-chain” actions (as opposed to “off-chain” actions). This means that a defendant who does not take any action on the blockchain (e.g. “calling a function”)Footnote 108 would not be liable for interference. This to some extent mirrors the fact that, if a defendant does not interact with or physically damage a chattel, he would not be liable in the chattel torts (unless there is a total impairment of use).
This in essence could be seen as underpinned by a “fair warning” rationale that preserves the liberty of the defendant: one should interact with the blockchain at one’s own risk, but there is no liability if one does not interact with the blockchain. Similarly, with the chattel torts, the message is that one interacts with chattels at one’s own risk, but there is no liability if one does not interact with chattels (unless there is a total impairment of use).
To approximate the chattel tort position further, this on-chain interference requirement could be imposed in relation to partial impairments of use, but not for total impairments of use.
However, there are two problems with using an on-chain interference requirement as a proxy. First, this requirement does not actually provide fair warning. Any on-chain function (when executed) may be a triggering condition for some other digital asset being burned or frozen.Footnote 109 To prevent this outcome, the defendant would need to search all of the smart contracts that exist and ensure that no digital asset would be burned or frozen as a result of executing/calling the intended function. Indeed, the analogy between “interacting with a physical asset” and “interacting with a blockchain” is a loose one. In the physical asset context, it is reasonably expected that one is supposed to “keep off” a physical asset. This is because (1) the general expectation is that someone may (or is likely to) own it and (2) one can avoid interacting with it because it has visible boundaries, meaning that a defendant can keep off it without expending much mental energy.Footnote 110 In contrast, interacting with a blockchain (even intentionally) does not (and ought not to) give rise to the expectation of “keeping off any digital assets”, because (1) a digital asset has no visible boundaries and (2) it is difficult to avoid conclusively the outcome of a digital asset being burned or frozen as a result of an on-chain interaction.
Second, the on-chain interference requirement would result in an under-inclusive rule that protects claimants insufficiently. For example, it would exclude a distributed denial of service (DDoS) attack by a defendant (which can be an off-chain activity) that prevents the claimant from being able to access his digital asset. This would happen, for example, if the claimant’s private key is stored on a particular website and/or mobile application that is the subject of the DDoS attack.Footnote 111 Another example would be where there is a DDoS attack on blockchain network nodes that prevents the claimant from being able to access his asset for a substantial period of time.Footnote 112 There could also be a DDoS attack on the relevant application programming interface (API) that connects the front-end application or website (used by the claimant to access his digital assets) with the blockchain back-end architecture, meaning that the claimant would be unable to access his digital asset through the application or website.
Another possibility would to be to impose a directness requirement for digital asset interferences, by analogy with the directness requirement in trespass to goods.
However, the analogy breaks down on a very fundamental level. With physical assets, we know what should be directly caused in the trespass context (i.e. the physical interference: physical contact or physical damage). In contrast, with digital assets, we do not know what should be directly caused (i.e. the impairment of use, the function call, etc.). Thus, imposing a directness requirement merely begs the question of what proxy we should use as an equivalent of physical interference, because we need to know what needs to be directly caused before the analogy with trespass can stand.
Nonetheless, could we use a directness requirement to constrain the scope of interference, even though we do not know what must be directly caused? It is suggested that a directness requirement is too vague. Directness usually means sufficient proximity in relation to the (1) time taken between the defendant’s action and the relevant consequence and (2) number of events between the action and the consequence, and also means that (3) the original action carries a high degree of influence in generating the relevant consequence.Footnote 113 However, these three elements are very open-ended and are difficult to apply to digital assets since it is difficult to generate a determinate result purely from applying these criteria. For example, it is difficult to know the boundaries of each of the criteriaFootnote 114 and the weighting of each factor, and this creates room for judges to be able to manipulate instrumentally the criteria to reach a desired result, at the expense of the law’s predictability and consistency.
Another proxy may be one based on the direct linguistic equivalent of “(intentional) physical contact”: namely, “(intentional) digital contact”. However, if “digital contact” (i.e. interacting with a digital device or system) were to be the equivalent threshold, there would simply be no fair warning to potential defendants (especially since conversion is a strict liability tort).
This contrasts with the position as regards physical assets, because physical assets have a boundary (and so avoiding intentional physical contact or damage is relatively easy). It is (relatively) not a big ask to require a person not to make contact intentionally with physical objects in his proximity, not to perform intentional acts that lead to damage to physical objects and not to perform intentional acts that lead to someone being totally excluded from using a physical object. Given these constraints, it may be argued that imposing strict liability provides a substantial degree of fair warning to defendants.Footnote 115
By contrast, in the digital asset context, people intentionally interact with functions on-chain and update data off-chain, in a way that may cause damage, destruction, or exclusion of access to digital assets without them knowing. It is difficult to know when one's conduct will cause such consequences.
For example, in the context of the prediction markets, someone might deploy a smart contract for the purpose of a sports bet and designate a website as the source of authority for the final score. Funds (in the form of digital assets) would be locked up in that smart contract (and could, for example, be jointly owned by all parties to the bet) and subsequently distributed to the person who wins the bet. The smart contract can designate any website as the “ultimate source of information” for the final score and the problem arises where such a website misreports the relevant score, causing the tokens to be distributed to the wrong person.
In this case, there is potential liability for causing the “diversion” or “misappropriation” of a digital asset. This liability could attach to the operator of any website that shows sports scores. It is difficult for the operator of any such website to know whether the information displayed on their website is being used for the purposes of a smart contract oracle. Extrapolating further, the same issues would apply to prediction markets more generally and, importantly, in the context of financial derivatives.
When coupled with strict liability, the requirement of digital contact would make people hypervigilant, would lead to a waste of people’s mental energy in thinking about whether they might be liable and might cause them to take preventive or defensive action. As such, their liberty (and economically useful activity) would be stifled. This position is similar to that in respect of pure economic loss, where a high mental requirement is imposed.Footnote 116 The law does not impose strict liability for causing pure economic loss, because doing so would stifle ordinary activity and make people hypervigilant.Footnote 117 Indeed, the ethos of the blockchain as an open-source environment where people are encouraged to experiment with codeFootnote 118 should be respected: imposing strict liability in the blockchain environment would run directly counter to such an ethos.
B. Blockchain Environment/Policy (v Physical Environment/Policy)
The blockchain environment is very different from the physical environment and this means that the policies that are relevant in the blockchain world produce a different normative balance as compared to those in the physical world. This means that the scope of defences available (as well as the scope of prohibited actions) in respect of digital asset interferences are likely to be very different to that in respect of physical assets.
For example, there can be kill switches and other coded permissions that are given to people so that they can (e.g.) burn or freeze an asset. If such people burn or freeze an asset with the intention of doing so, this would be an intentional action that directly leads to C’s use being impaired or destroyed (as he would not be able to transfer the asset and (in the case of burning) not be able to access the asset as well). Such situations may arise, for example, if there has been a bug or hack (or a suspected bug or hack) and the developer/administrator exercises the kill switch to investigate what has been happening with the code, with the effect that the claimant is no longer able to use or access his asset at all.Footnote 119 This would constitute the equivalent of a physical conversion,Footnote 120 and there is no defence to conversion that applies here. The consent defence would not apply in many (or most) instances since claimants in many (or most) instances would not even be aware of such a kill switch or permission,Footnote 121 and so there is no express or implied consent.Footnote 122 Also, even though the defendant may argue that the claimant opted into the blockchain world (and his particular protocol) and thus consented to its rules (or “logic”), this would not succeed. This is because such “logic consent” does not involve actual consent (whether express or implied) or perhaps even hypothetical consent (i.e. where the claimant would have consented if he was made aware of the effect of the protocol rules).
However, there is a strong argument that this is the wrong normative resultFootnote 123 and that a defence ought to be available in (at least some) such circumstances. This is because the kill switch/permission was constructed as part of the blockchain environment. It was deliberately created (i.e. the effect was intended), as opposed to being an accidental consequence of bad programming.
A particular blockchain environment can be constructed in a multitude of ways and one can create the rules that govern such a blockchain environment/application. This stands in contrast with the physical environment where the physical laws are relatively fixed/immutable. As such, if a blockchain developer decides to design a blockchain that includes a kill switch that is intended to be used in circumstances where the assets are in danger of imminent misappropriation or destruction by a hacker or in danger of being destroyed by an unintended bug, affording no defence to the developer where those precise circumstances exist (despite the lack of consent from the claimant) may be thought to be unfair. This is because it frustrates the very purpose of the developer’s deliberate design choice to add in a kill switch to protect the integrity of the blockchain.Footnote 124
Also, a blockchain environment is an opt-in environment (in the sense that a person can choose whether to engage with it or not), unlike the physical environment (which every person has no choice but to engage with). If one chooses to enter a blockchain environment, it would seem fair to suggest that, in certain circumstances within people’s reasonable expectations,Footnote 125 those with kill-switch/burn permissions should be allowed to exercise their power to transfer/freeze assets as intended by the design of the blockchain.
However, a claimant is not reasonably expected to look out for his blockchain environment in the same way that he is expected to look out for his physical environment. Looking out for one’s physical environment merely requires one to be (visually and kinaesthetically) attentive to one’s surroundings, which is already habitual for most people and thus does not take much effort. In contrast, looking out for one’s blockchain environment carries vastly higher information costs, as it requires a detailed understanding of code (which the vast majority of people do not have) as well as the ways in which it could malfunction. Indeed, even programmers cannot anticipate all bugs/loopholes in the code, so it would be extreme to suggest that an average user of the blockchain should be expected to do so. Thus, if the particular use of a kill switch is outside a claimant’s reasonable expectations, a defendant in general ought not to be afforded a defence.
Overall, it is clear that the threshold of fair warning and the relevant contextual considerations are very different across physical and digital assets, which means one would expect a substantially different scope of liability in respect of interferences with the latter.
VI. Argument 2: Chattel Tort Rules Themselves Are Unsatisfactory and Arguably Too Harsh
The law surrounding chattel torts is unsatisfactory. The rules themselves are messy and needlessly complex, in that they are weighed down by unnecessary conceptual baggage and shrouded in vague language. Also, the strict liability of the chattel torts is arguably too harsh as it gives rise to problems for innocent defendants. Therefore, applying the same rules to digital assets would lead to the same undesirable features being replicated in the digital asset context.
Three features will be explored: (1) the “right to immediate possession” concept, (2) the lack of a universal definition of conversion and vague formulations of the tort and (3) the strict liability nature of the chattel torts.
A. Right to Immediate Possession
The right to immediate possession concept is highly problematic and should not be applied in the digital asset context.Footnote 126 It is used to determine the outer limits of whether a person has title to sue in conversion,Footnote 127 but the language of right to immediate possession is fundamentally vague and does not provide much guidance. It is unhelpful insofar as “right to immediate possession” is synonymous with “right to sue for interference with possession” and it is also unhelpful insofar as a person with title to the good already has a right to possess the good.Footnote 128 In this sense, the notion of a right to immediate possession does not provide much further guidance on what is required to have title to sue for conversion. It does not easily map onto the threshold of “you must have titleFootnote 129 but not have granted a chattel lease or pledge” and many judges have made mistakes (e.g. in holding that a mere contractual right to possession confers title to sue).Footnote 130
This vague language of right to immediate possession is problematic because it obscures the issue of whether the chattel lease is a derivative interest, as well as the numerus clausus debate that informs it.Footnote 131 These issues are fundamental and need to be clarified, because the general principle across property law is that a person with title can sue for interference unless he has granted a derivative interest. Yet, these considerations are not confronted in the conversion context and are hidden under the concept of the right to immediate possession. At the same time, the technical language gives it the appearance of legitimacy despite it merely being a conclusory label: “a right to immediate possession” describes the default rights of a person who has title (insofar as he has a right to exclude), but is vague and unhelpful as a test for who has the right to sue for interference. This gives judges room to reach a desired conclusion without transparent reasoning to justify it.
If the right to immediate possession concept were to be applied in the digital asset context, this conceptual confusion (as well as room for opaque reasoning) will be replicated, and the fundamental numerus clausus and fragmentation of titleFootnote 132 issues will not be clarified let alone confronted.
B. Definitions and Formulations of Conversion
Furthermore, there is no universal definition of conversion and the current formulations of conversion are vague. Indeed, Lord Nicholls in Kuwait Airways Corpn. v Iraqi Airways Co. (Nos. 4 and 5) Footnote 133 stated that it is “well nigh impossible” to define the tort,Footnote 134 and Selvam J. also noted that conversion is “too elusive to be expressed in words”.Footnote 135
The lack of a clear formulation of the tort makes it difficult for people to know the normative threshold for interference, insofar as it requires them to slice through a layer of conceptual baggage to discern what the general normative threshold is.
For example, conversion has been referred to as a denial of the claimant’s title to the chattelFootnote 136 by a defendant’s assertion of title over the claimant’s chattel.Footnote 137 This formulation does not provide much practical guidance, as it does not answer the very questions of (1) what acts constitute a denial of title and (2) what acts constitute an assertion of title by the defendant. Similarly, conversion has been described as an “intentional act or dealing with goods” that is “inconsistent with or repugnant to the rights of the owner”.Footnote 138 This again does not provide much practical guidance, because it fails to answer the very question of when the act becomes “inconsistent with” or “repugnant to” the rights of the owner.Footnote 139 It is far from clear that formulations like “denial and assertion of title” or “inconsistency” or “repugnancy” map onto the substantive threshold for conversion set out in Section IV.Footnote 140
Such formulations, if interpreted at face value, can produce many different results. It is difficult to infer the general threshold for interference from such formulations/definitions and it is necessary therefore to slice through a layer of conceptual baggage in order to discern what the general threshold is. This causes confusion and uncertainty: indeed, the authors of The Law of Personal Property note that “much of the difficulty in conversion lies in estimating the required seriousness of the defendant’s interference”,Footnote 141 and similarly Douglas notes that it is an “almost impossible task for a lawyer to advise a client on the merits of a possible claim”Footnote 142 outside certain well-established categories of conversion. These problems will be replicated in the digital asset context, as judges will not be able to discern with ease what the threshold requirement for interference is and may reach arbitrary decisions because of the lack of guidance provided by the existing formulations of conversion.
Also, the fundamental purpose of vagueness is to provide the necessary flexibility to respond to context, to avoid running the risk of making the formulation overly precise (which causes over-inclusiveness and under-inclusiveness).Footnote 143 However, the vagueness that exists in the chattel torts goes beyond what is necessary to respond to context. One can be a lot more precise than this, for example, by adopting a formulation of conversion that makes it reasonably clear what is required: Douglas, for example, suggests that conversion should be defined as an “intentional exercise of exclusive control” over another’s chattel.Footnote 144 As with the right to immediate possession, the existence of vague rules also creates room for judges to bendFootnote 145 doctrine to reach the right normative result, which compromises the transparency of the judge’s reasoning. This ought not to carry forward into the digital asset context.
C. Strict Liability
Furthermore, the strict liability nature of the chattel torts is arguably too harsh on innocent defendants, as it holds them liable for honest but reasonable mistakes.Footnote 146 For example, if an asset is sent to the defendant and he sends it to someone else under the (honest and reasonable) mistaken belief that the asset is his, he commits conversion and is liable to pay the full value of the goods to the defendant.Footnote 147 This can be seen as an unfair result, because it often defeats reasonable expectations of the defendant, given that there are many situations where such a defendant may very reasonably believe that he owns the relevant good.Footnote 148 Also, it becomes very hard for people to plan their activities in such a way that they can predict the legal consequences of those activities and they would either (1) need to expend time and cost to verify whether the asset is owned by someone else (in order to avoid the risk of liability) and often never find out the answer or (2) simply refrain from acting (or attempt to avoid getting themselves into situations where goods may be owned by a third party). Insofar as this result is thought to be unfair,Footnote 149 it should not be replicated in the digital asset context.Footnote 150
For example, a centralised digital asset exchange could receive a cryptocurrency from a client (especially if it has done the relevant AML/KYCFootnote 151 checks) into its own address where the client did not have the best title.Footnote 152 In this case, the exchange would receive the client’s (inferior) titleFootnote 153 and either hold it on trust or hold it outright subject to a contractual obligation to return an equivalent quantity of cryptocurrency.Footnote 154 If the exchange later uses the cryptocurrency for proprietary trading and executes an on-chain transfer of the cryptocurrency to its counterparty (without knowing of the defect in title),Footnote 155 this would prima facie constitute a conversion in the absence of a bona fide purchase defenceFootnote 156 and the true ownerFootnote 157 could bring an action against the exchange. This can be seen as an undesirable result as it causes problems for innocent exchanges: they may end up being liable in conversion and thus take defensive measures to avoid this risk, even if such measures cause a wasteful depletion (or otherwise inefficient use) of their assets.Footnote 158
VII. “Digital Trespass and Conversion” Cases in Other Jurisdictions
At this point, one may nonetheless argue that the existence of “intangible trespass and conversion” cases in other jurisdictions means that extending the chattel torts to digital assets does not pose a problem. It is argued that this conclusion is mistaken, for two reasons.
First, most of the cases involving conversion and trespass to intangible assets do not involve an interference with digital assets: they mainly involve impairments of use of digital files or domain names.Footnote 159 It does not follow that digital assets should be covered by conversion and trespass.Footnote 160 There are policies specific to the digital asset environment that are not relevant in the context of digital files or domain names, such as the composability of the underlying blockchain environment. This leads to the wrong normative threshold being applied to digital assets, for example where (as mentioned earlier)Footnote 161 the consent defence in conversion and trespass does not take into account the fact that a defendant blockchain administrator ought to be able to impair use of a claimant’s digital assets in some situations where he needs to engage a kill switch and freeze the state of the blockchain in order to prevent a suspected bug or hack.
In any event, some of the conversion cases involving intangible assets can be recharacterised. For example, where the claim involves impairment of use of a domain name (e.g. where the defendant fraudulently persuades the domain name registrar to register a domain name in his favour), this is in essence a claim for pure economic loss. A domain name involves a contractual right against the service provider and a domain name by itself confers no title to any separate object of property (such as the underlying computer servers that perform the relevant operations in respect of the domain name). Here, there would be a remedy under the unlawful means tort.Footnote 162 It is suggested that, since there is no interference with any concrete object of property to which the claimant has title or a possessory interest,Footnote 163 domain name interferences should not by themselvesFootnote 164 fall within the scope of the property torts.
Second, although there are US cases that have applied the chattel torts to digital assets,Footnote 165 this does not change the fact that the two substantive arguments in this article still apply as far as English law is concerned. The fact that the chattel torts apply to digital assets in the USFootnote 166 does not remove the fact that the cases often involve complex factual scenarios that require nuanced balancing acts (where the English chattel torts are unsuited to dealing with). An example would be Shin v ICON Foundation,Footnote 167 where an unintended loophole in the code caused the claimant to mint extra tokens for free. The defendant administrators of the blockchain system froze the claimant’s tokens without his consent.Footnote 168 The system rules enabled such freezing to occur and the act of freezing was intended to reverse the effects of the loophole/bug in the code.
The conversion issue was not fully dealt with in the judgment given the nature of the application,Footnote 169 but a judge who is tasked with tackling the issue in full would need to deal with a complex and nuanced balancing act. One needs to balance (1) the fact that the defendant impaired the use of the claimant’s digital asset and (2) the fact that the claimant minted extra tokens in accordance with the rules of the system/program, against (3) the defendant’s desire to reverse the unintended effects of the program.
If we were to apply the English law chattel torts to the facts of Shin, we would again run into the risk of the physical asset threshold producing the wrong normative result in the digital asset context. Assuming there is a relevant interference (based on the fact that there was a complete impairment of use of the claimant’s digital assets), the consent defence would not be available to the defendant since the claimant did not consent to the freezing of the tokens. However, the consent defence does not take into account the fact that it may be justified for the defendant to exercise a kill switch especially in situations where there was an unintended loophole/bug in the code that was exploited by the claimant. Ultimately, after balancing the relevant considerations, a judge may reach the conclusion that, from a normative perspective, there should not be a defence, but the problem is that there is simply no doctrinal tool capable of accommodating these competing considerations. Having no defence to consider apart from consentFootnote 170 means that this crucial issue does not even end up being discussed.
VIII. Conclusion
This article has argued that, although there is a gap in protection in respect of digital assets, it should not be filled by extending the chattel torts. Physical and digital assets are too different in terms of their nature, behaviour, environment and underlying policies, and judges are substantially less familiar with digital assets as compared to physical assets. This means that the concepts that are used to resolve disputes in the physical asset space are either unhelpful or produce the wrong normative result in the digital asset space. This is compounded by the fact that chattel tort rules are unsatisfactory in that they are vague, needlessly complex and create problems for innocent defendants, and so applying the chattel torts to digital assets will lead to a replication of these mistakes in the digital asset context.
As such, the task of ascertaining the appropriate scope and method of protection in respect of digital asset interferences should be approached without the baggage of the chattel torts. Ideally, it should be approached afresh, so that the normative issues can be tackled from the ground up.
Open access
