Are Smart Contracts Real Contracts?

Alongside the rise of NFTs, smart contracts have received quite the fanfare over the past year. But, in truth, they’re a pretty old-school phenomenon. In fact, the first smart contracts had little to do with NFTs, blockchain, or the internet, and more to do with the humble vending machine. 

The first fully automated, coin-operated vending machines were introduced around 150 years ago to London railway stations and post offices, offering merchandise for sale in a convenient manner. When the requisite amount of money was inserted into the machine, the product — perhaps a postcard or postage stamp — would be dispensed automatically and a contract of sale automatically executed between seller and purchaser. Though the range of products available for purchase through vending machines has since evolved to encompass pretty much any commodity, from credit vouchers to NFTs, the automation of sales transactions continues today.

In 1996, more than a decade before the release of the original blockchain white paper by Satoshi Nakamoto, the computer scientist, cryptographer, and legal scholar Nick Szabo published an article in which he first used the phrase “smart contract.” At the time, he was completing a graduate degree in Law from The George Washington University Law School. In the text, Szabo defined a smart contract as “a set of promises, specified in digital form, including protocols within which the parties perform on these promises.” And he expressly linked smart contracts to vending machines, describing them as the “primitive ancestor of smart contracts.” 

But Szabo pushed the automation concept even further, suggesting that “[s]mart contracts [could] go beyond the vending machine in proposing to embed contracts in all sorts of property that is valuable and controlled by digital means.” In this way, he saw the potential of smart contracts to address the challenges of enforcing legally binding contracts. He also questioned the requirement for a judicial system to decide what physical steps are to be taken by an enforcement agency, if contractual arrangements were readily verifiable. For Szabo, contractual enforcement could become proactive, rather than reactive, if smart contracts were “embedded in the world.” 

But Szabo’s theory of smart contracts is just one approach to contract automation. In recent years, the global legal industry — from law firms to in-house legal operations — has largely embraced AI, machine learning, and automation technology as a way of reducing the time (and therefore costs) of more mundane, data-intensive tasks such as contract management and trial discovery. 

In 2018, real estate company CBRE UK conducted a survey of 100 London-based law firms, finding that 89% of respondents were already using AI or had imminent plans to do so. Today, the AI software market in the legal industry continues to expand, and is projected to grow to over 28% between 2021 and 2026. Keeping up with advances in technology has even been mandated by legal industry bodies. For example, in 2012 the American Bar Association amended part of its “competence” requirement for legal professionals, providing that:

“[In order] to maintain the requisite knowledge and skill, a lawyer should keep abreast of changes in the law and its practice, including the benefits and risks associated with relevant technology, engage in continuing study and education and comply with all continuing legal education requirements to which the lawyer is subject.”

The range of smart contracts has also diversified over time. These days, the degree to which a smart contract is automated and/or decentralized is more akin to a sliding scale than an either/or. A case in point is an online agreement like a clickwrap agreement. Consider, for example, Uber’s Terms of Service or “T&Cs.” When a user downloads the Uber app, a contract is formed between the user and Uber. The T&Cs spell out the rights and obligations of the user and Uber. While the form of the contract is digital, its substance is written language. 

The enforceability of Uber’s T&Cs was considered in the US case of Meyer v Uber Technologies Inc., No. 16-2750 (2017). In this case, the plaintiff Spencer Meyer brought an antitrust conspiracy claim against Uber’s then CEO Travis Kalanick and the Uber company, alleging that the algorithm used by Uber’s app allowed drivers to fix prices. In response, Kalanick and Uber filed a motion to dismiss the court proceeding and resolve the dispute by arbitration instead. As the T&Cs included a term which required users to enter into arbitration proceedings in the event of a dispute with Uber, Kalanick and Uber argued that Meyer was contractually required to enter into arbitration proceedings, rather than court proceedings in the first instance. 

One of the key legal issues to arise in this case was whether Meyer and Uber had entered into a legally enforceable contract. Meyer argued that he had not been given reasonable notice of the T&Cs when he registered as a user and downloaded the Uber app, therefore a legally enforceable contract had not been formed. But on appeal, the court disagreed with this argument and found in favour of Uber. It held that a reasonable smartphone user would have constructive notice of the T&Cs, regardless of whether they had bothered to read them. While this finding is not legally binding outside of the US, courts in other jurisdictions have taken a similar approach. 

Another genre of smart contracts is “data-oriented contracting” and “computable contracts.” These were first identified by legal scholar Harry Surden, whose theories built upon Szabo’s smart contract concept, as well as cryptographer Ian Grigg’s “Ricardian Contract” which defined a type of value that could be read by a computer. In a 2012 article titled “Computable Contracts,” Surden argued that it was possible “to represent contractual obligations in forms other than ordinary language,” including computer-processable data. According to this approach, based on “data-oriented contracting,” parties can make a contract term readable by a computer. One such example is foreign currency derivative contracts, which include computer-readable data records to manage financial risk. 

With computable contracts, Surden took the possibilities of smart contracts up a notch. In a similar vein to Szabo’s concept of self-enforceable protocols, Surden suggested that computable contracts could enable a computer to compute using data analysis. Nowadays, the floodgates for this sub-genre of smart contracts have well and truly opened, propelled by the Ethereum blockchain, which allows anyone to build smart contracts on top of its platform.

But Surden also acknowledged that computable contracts have limits, and are not suitable for all contractual arrangements. For example, computable contracts are not suitable for arrangements that rely on abstract concepts or require discretion by the parties. For example, a party may agree to try to achieve something, and the contractual term may oblige said party to take reasonable efforts to do so. 

Computable contracts should also be distinguished from the efforts being made by the legal industry to implement AI and machine learning tools. Following Surden, a number of theorists have pointed out the limitations of computable contracts in only being able to replace contractual terms which are specific and objective. In their seminal book Blockchain and the Law (2018), Primavera de Filippi and Aaron Wright made the case for smart contracts, whilst also restating the importance of leaving some contracts open-ended, owing to the flexibility afforded by subjectivity.¹ 

Yet even now, the degree to which smart contracts are legally enforceable as contracts, and might even replace legally enforceable contracts, remains the province of theory and is yet to be tested properly in a court of law. 

A number of blockchain-based art projects have emerged in recent years that elaborate such theory. Consider the work of OG crypto artist Rhea Myers for example. In the early days of Ethereum, Myers created Is Art 2014/5, an Ethereum smart contract that contained the assertion that the work either “is” or “is not” art. A web page connected to the Ethereum network would display the current state of the assertion and anyone with access to the smart contract could toggle that contract’s state. For Myers, both 1960s Conceptual Art and mid-90s net.art movements serve as useful reference points in thinking together blockchain and smart contracts. In her documentation, the artist writes:

“These art movements stood in critical tension with the systems of communication, law and commerce of their eras. Each treated rootless information, whether about sense data or network messages, as the critical subject of art and a new potential artworld. Their promise and their eventual recuperation by the existing artworld chimes with the historical experience of the blockchain.”

1960s and 70s Conceptual Art is peppered with reflections on the social, economic, and legal conditions governing art transactions.² While Seth Siegelaub and Robert Projansky’s model contract for artwork sales, The Artist’s Reserved Rights Transfer and Sale Agreement (1971), is the most commonly referenced legal document-cum-conceptual artwork, another slightly earlier example is Daniel Buren’s Avertissement, written by Buren together with lawyer Michael Claura in 1968. Roughly translating as “Warning” in English, Avertissement has been used by Buren as the contract of sale for his artworks and, simultaneously, as a certificate of authenticity. 

The relationship between the authenticity of an artwork and its contract of sale has also been a source of enquiry for Myers. Certificate of Inauthenticity (2020), a somewhat tongue-in-cheek artwork in the medium of Ethereum’s ERC-721 smart contract “standard,” points to a digital image of a “Certificate” of a rare artwork that “has not been modelled, instantiated, or presented by Rhea Myers”. The certificate purports to validity only so long as the certificate is,

“organized by, on behalf of, or with the full knowledge and authorization of the party or parties that …control the smart contract or contracts that are currently the owner0f() the ERC-721 token having the URI of this image as the “image” field of its tokenURI() metadata...”

In adopting a numerical style and mode of expression commonly used in written language contracts, Myers foregrounds one of the crucial characteristics of legally binding contracts — certainty. At the same time, she calls attention to the truth claims of an NFT: 

“It is difficult to represent off-chain truth on-chain, however. It is prohibitively expensive to put bitmapped digital images on-chain. And it is impossible to make sure that a digital image is represented uniquely even within the same smart contract on a single blockchain. This means that we must trust the claim that the image represented by each ERC-721 token is what its creator claims.”

Does this mean that smart contract standards are simply not “smart” enough to support rare artworks? Would a written language contract actually be more useful? And what exactly is the legal status of the millions — if not billions — of blockchain transactions that involve artworks minted using ERC-721? 

Last year, the artist Nancy Baker Cahill released a series of augmented reality artworks-turned-NFTs titled Contract Killers (2021). As well as driving much-needed discussion around the suitability of the ERC-721 standard for artworks, Cahill also used the vocabulary of smart contracts to illustrate the breakdown of social contracts. 

In one piece, two hands come together in front of City Hall as if to shake hands, only to dissolve before our eyes, “to underscore the dissolution of policies that force us to recognise that we belong to each other.” In another, they come together in front of the Hall of Justice, “to address the profiteering and injustices of the prison industrial complex, court system, and all the beneficiaries and victims therein;” and in the third image, they do so over a pile of virtual cash, “to point to the great inequities of late-stage capitalism”. 

A key component of Contract Killers is the accompanying written language contract between Cahill and the purchaser, drafted by the artist together with lawyer Sarah Conley Odenkirk. The contract includes a number of terms central to the NFT as a concept, including the ownership rights of the purchaser, resale royalty requirements, and a guarantee by the artist that only one of the three NFTs will be minted. Like Siegelaub and Buren before her, Cahill wanted to implement a robust, legally binding contract for the sale of her artworks, and wanted to share the contract with her public.

In Odenkirk’s paper, published to coincide with the project, she explained the rationale for an off-chain, written language contract as “the [NFT] marketplace is currently lacking in protocols, rules, remedies or any formal structure that would protect NFT owners or creators against any number of eventualities.” Moreover, the phrase “smart contracts” is “false and misleading” as “no actual agreements are included [in NFTs].” Indeed, for any trained lawyer familiarizing themselves with the behavior of crypto communities, the reliance on social contracts over legal contracts, can be a shock to the system. 

But does this mean that the “smart contract” needs to be abandoned?

There is so much to be learned, including by lawyers, from the fascinating history and theory of smart contracts. Granted, today’s ERC-721 has its limitations, but so too do written language contracts, which are all too often burdened by a lack of access to justice. Perhaps this is where artists and lawyers can continue to work collaboratively — on a new art of the smart contract.