Skip to content

#362 Tornado Cash: The developer as co-perpetrator of money laundering

16 May 2024

I have been eagerly anticipating the judgment on Tornado Cash in The Netherlands, due to its significant implications on legal interpretations of money laundering. The case against Pertsev, one of the creators of the code, focuses on the accusation of co-perpetrating the concealment of the criminal origins of specific Ethereum transactions.

The central questions are Pertsev’s exact actions, his awareness of the criminal origins of these transactions, and the extent of this knowledge relating to the criminal origin in relation to his concrete acts. Legal questions also arise regarding when one must be aware of the criminal origins of goods, how specific this knowledge must be, and the relationship in time of this knowledge to one’s actions. From my perspective, the judgment fails to address these questions adequately, indicating that it is ripe for appeal.

While the judgment covers multiple facets, this blog will concentrate on the legal issues described above.

Understanding Tornado Cash

The court details how Tornado Cash operates. Launched in August 2019, Tornado Cash is a privacy tool that allows users to deposit and withdraw cryptocurrency anonymously, aiming to leave no trace on the Ethereum blockchain. Essentially, it severs the links between deposits and withdrawals. For a more detailed explanation, I refer to the judgment.

The role of Pertsev in Tornado Cash

The court asserts that Pertsev, alongside co-suspects Roman Storm and Roman Semenov—who are under prosecution in the US—are the creators of Tornado Cash. They engineered the system to function autonomously and immutably. Since the system’s basic operations were unalterable from the outset, the court holds these individuals are and remain accountable for its functionality.

In December 2020, the governance of Tornado Cash was transferred to a Decentralized Autonomous Organization (DAO), meaning its governance structure was by then overseen by the community. However, the court concludes that the foundational operations of the system were unchangeable and, therefore, the original setup by the founders remains a critical factor.

Co-perpetration Concerns

The court’s reasoning on co-perpetration seems flawed to me. It suggests that Tornado Cash, let’s call it a digital platform or system, independently enacted the acts of hiding and concealing cryptocurrency with a criminal origin. This implies that a system can engage in criminal behaviour on its own—an unusual and potentially problematic interpretation, as it shifts the criminal conduct itself from individuals to systems.

According to our Criminal Code, the act of hiding and concealing should refer to actions that are ultimately performed by humans and, in certain situations, can be attributed to legal entities; a system itself cannot engage in criminal behaviour. It is for instance not a car that creates an accident, but the person driving the car. Tornado Cash cannot do anything in itself until a person deposits the cryptocurrency, the act of a human is needed.

This is exactly where the court’s reasoning fails, in my opinion. The court incorrectly states that the system should not be seen merely as a tool to facilitate criminal conduct, but rather asserts that the tool itself directly engages in criminal conduct. This is a crucial part of the court’s reasoning to qualify Pertsev as a co-perpetrator of money laundering.

Co-perpetration occurs when two or more persons engage in close and conscious collaboration in committing the criminal offence. Because there has been no interaction whatsoever between the ‘criminals’ who deposited the crypto, the court came up with this contrived reasoning of Tornado Cash being ‘the one’ hiding and concealing, and the “founding fathers” are held responsible for the actions of the tool. This could have significant implications for software developers if tools are perceived as independently capable of criminal actions, making the developers responsible for the tool’s ‘actions’.

Questioning criminal intent

Another aspect open to criticism is how the court motivates Pertsev’s criminal intent. Pertsev was indicted for his alleged role in concealing the origins of 37 transactions worth about $1.2 billion. The judgment does not clarify when these transactions occurred or if Pertsev was even aware of them, which seems unlikely given the system’s design. The court’s assumption that the general knowledge that mixers could be used for criminal activities was enough to establish Pertsev’s intent for these specific transactions is questionable.

The court argues it is common knowledge that mixers are used by criminals, citing a report by the Financial Action Task Force and the classification by the Financial Intelligence Unit in The Netherlands dated August 15, 2017, that the use of a mixer is a money laundering typology.

However, not all money laundering typologies are recognized as facts of common knowledge. The rationale for attributing criminal intent heavily relies on this assumption, which is quite far-reaching. Does foreseeing that your software might be used by criminals make you criminal liable for their actions? And what did Pertsev know at that moment in time, as the assumed common knowledge in 2019 is highly debatable?

The case file might indicate Pertsev was aware of criminal usage of Tornado Cash at some point, but this does not necessarily relate to his actions at the time he set up the system or to the timing of the specific transactions. And what were his actions or could his actions have been at the time of the transactions given the fact that the basic system was unstoppable? The court remains silent on these points, leaving significant legal questions about the relation between knowledge and acts unanswered.

I believe the judgment for Pertsev’s criminal intent was poorly motivated, and the legal intricacies remain underexplored. An appeal will be necessary to address these issues thoroughly.

Gepubliceerd door onze specialist:

J.N. (Judith) de Boer