The past decade has seen a reinvigorated interest in the concept of trust, primarily driven by the onset of the global financial crisis in 2008, which has been commonly attributed to the failure of trusted institutions such as banks and other financial institutions. Blockchain technology has emerged as a potential solution to address this erosion of trust, as it allegedly eliminates the need for trust between parties. Indeed, the underlying premise of blockchain technology is that users subject themselves to the authority of a technological system that operates according to predefined code-based rules, rather than to the authority and discretionary power of centralised institutions which are deemed untrustworthy. This has led to the common misconception of seeing blockchain technology as a ‘trustless’ system or even as a ‘trust machine’. We argue that this misconception is grounded on an incorrect understanding of the notions of ‘trust’ and a neglect for the related concept of ‘confidence’ and that—in light of its technical underpinnings—blockchain technology should rather be regarded as a ‘confidence machine’.
Trust depends on personal vulnerability and risk-taking. It exists when one party (the trustor) voluntarily decides—under a condition of uncertainty—to rely on another (the trustee) for the achievement of a particular task, based on the belief that the latter will perform the task in line with the expectations of the former. Trust is beneficial because it enables the trustor to economise resources by (a) delegating to a third party the performance of a task, and (b) reducing the level of direct involvement needed to ensure the proper performance of that task. However, the flipside of trust is that the trustee has the ability to act against the interests of the trustor.
Luhmann further elaborates on the distinctive properties of trust by distinguishing it from the feeling of confidence (Luhmann 2000). Confidence, as opposed to trust, does not require an individual to put herself into a vulnerable position because it does not operate under a condition of uncertainty. It depends on internalised expectations deriving from knowledge or past experience, and, as such, does not presuppose an acknowledgement of risk, but rather an attitude of assurance. Confidence derives from the predictability of future events: it emerges when an individual believes that the person or system she interacts with does not have the agency to betray her expectations.
Of course, the two terms are inherently interconnected. Confidence in a system ultimately depends on the level of trust or confidence that one has in the actors or institutions involved in higher-order systems. As Giddens explains, much of the confidence we experience in our daily activities only subsists because of the trust we have in a variety of expert systems (eg the legal system, professional guilds, the scientific community, etc) which we believe provide the necessary ‘guarantees’ for us to build expectations on matters which we do not have the ability to exhaustively verify on our own (Giddens 1990).
It is commonly understood that the raison d’etre of a blockchain-based system is that it does not require trusted third parties. As described by blockchain advocate Antonopoulos, blockchain technology enables a ‘shift from trusting people to trusting math’ (Antonopoulos 2014). Werbach reiterates this point by defining blockchain technology as an enabler of ‘trustless trust’, where transactional security is achieved via reliance on deterministic computation (Werbach 2018). But what is trust actually replaced by?
We argue that blockchain-based systems are intended to produce ‘confidence’ in a particular system—not by eliminating trust altogether, but rather by maximizing the degree of confidence in the system as a means to indirectly reduce the need for trust. Indeed, to the extent that one can understand the code of a particular piece of software, it becomes (theoretically) possible to predict the output of that software, for any given input. Hence, the higher the predictability of the software code is, the higher is the confidence in the system and the lower is the need for trust in the developers and/or operators of that technological system.
Confidence in a blockchain-based system is achieved through a combination of multiple elements. First, there is confidence in the mathematical rigour of the hashing algorithm, especially with regard to the cryptographic primitives that constitute the underlying foundation of a blockchain (eg public-private key cryptography, hashing functions, etc). Second, blockchain-based networks generate confidence in the economic incentives and game theoretical schemes that govern the network, grounded on the premise that miners will always act in such a way as to maximize their financial rewards. Yet, because miners are ultimately controlled by people, and might therefore be bribed or corrupted, additional guarantees have been introduced into these systems in order to further reduce the need to trust any individual miner. On the one hand, the consensus algorithm of most blockchain-based networks (eg Proof of Work or Proof of Stake) is intended to distribute trust among a large variety of miners, thereby reducing the risk of individual opportunism. On the other hand, because all participating nodes (such as miners and validators) hold a copy of the blockchain, they can always verify that every recorded transaction is valid and legitimate. Hence, anyone interacting with a blockchain may have a high level of confidence that it will operate as planned, even if they do not know (and therefore do not trust) the parties operating or maintaining the network.
Yet, despite the increased confidence that blockchain technology potentially provides, the existence of uncertainty and risk becomes apparent upon closer inspection. Although a blockchain-based network might appear, at first glance, to operate in a deterministic and self-contained manner (ie independently of the influence of third parties), the reality is that these networks are hybrid systems made up of both technical and social components. The governance of a blockchain is determined not only by code, but also by a variety of actors, including miners, validators, programmers, cryptocurrency and token holders, end-users, and, to a lesser extent, regulators. Having confidence in the system ultimately means trusting the whole assemblage of actors associated with that network. Thus, blockchain technology may well reduce the need to trust any one of these individual actors, however, it does not eradicate the need for trust altogether. The technology displaces trust in the technological artefacts that underpin a blockchain-based system, and shifts it towards the network of actors that contribute to operating and maintaining the system.
In order to prevent these actors from operating in an untrustworthy manner and thereby undermining confidence in the system, good governance becomes a central issue for public blockchains. In this recently published article, we reflect on the limitations of on-chain governance in building trust and confidence and seek to draw lessons from earlier attempts at constitutionalizing internet governance and polycentric governance for establishing the Rule of Law in a decentralized system.
Primavera De Filippi is a Permanent Researcher at the National Center of Scientific Research in Paris and a Faculty Associate at the Berkman Klein Center for Internet & Society at Harvard University.
Morshed Mannan is a PhD Candidate at Leiden University.
Wessel Reijers is a Max Weber Fellow 2018-2020 at the European University Institute.