IEEE Technology and Society Magazine - March 2020 - 74

Ethereum's Decentralized Autonomous Organisation
(DAO) has already provided an infamous example of what
can go wrong when moving trust from human institutions
to algorithms. As recounted by Tomaso Aste and colleagues [16], an unknown party exploited a bug in the
DAO's contract to make off with 3.6 million ether (worth
around US$50 million at the time). Aste and colleagues go
on make the observation that "The DAO was a failure only
from the viewpoint of its investors. From a technical perspective it had worked seamlessly" [16, p. 22]. That is, the
DAO did exactly what it was told to do by its programmers
in response to the input that its users gave it. The end
result nonetheless disappointed everyone involved (except,
presumably, the party who ended up with the ether).
Karen Levy characterizes smart contracts as "booksmart not street-smart" for these reasons [17]. She argues
that traditional contracts are not just technical instruments for enforcing an agreement, but also have a social
function in the way they set the expectations and maintain the relationships of each party to the contract. A
street-smart contract needs to be sure that the computer
is being complexified such that it can handle the world,
not the world being simplified such that it can be handled
by a computer.
Even if smart contracts are not sufficiently streetsmart to replace all of the functions of contract law, it
remains reasonable to ask if there are specific application domains in which smart contracts are able to solve
a relatively constrained problem. The finance industry,
in particular, has identified a number of tasks that at
least some think may be suitable for automation via a
blockchain [18], and other researchers are investigating
medical applications [19] and various schemes for managing identity and ownership [20]. This article, however,
does not have space to critique each of these applications individually.
People agreeing to smart contracts may also appreciate constraining the problem. As noted earlier, smart
contract languages resemble general purpose programming languages, so how is an ordinary person to know
what he or she is actually agreeing to? Digital rights management applied so-called rights expression languages
that, while not especially readable in their raw form,
expressed licenses according to a model that could be
mapped to concepts that ordinary people understood.
A few researchers have begun to address this issue by
designing "templates" from which an ordinary person can
generate a contract that he or she understands. In one
proposal, people would obtain a boilerplate contract from
some reputable organization and fill in parameters
according to their needs [21]. In another, proto-contracts
would be written as a series of rules relating one concept
to another, similar to the way in which rights expression
languages work, such that these could be converted into

executable contract code [22]. What is arguably one of the
more successful applications of rights management
ideas - the Creative Commons licenses - work in a similar way, by providing a set of choices easily understood
by ordinary people, and mapping these choices to the
more complex code used by lawyers.
Digital rights management workers considered two
main ways in which intractable problems like fair use might
be addressed: by permitting use of a "circumvention
device" to remove the rights management under certain
circumstances, such as for archival by libraries or for use
by the disabled; or by building an override mechanism into
the rights management system in the first place. The former approach was enshrined in many laws, but rarely if
ever exercised, while the latter remained only a theoretical
proposal. Circumvention devices or override mechanisms
might be constructed for smart contracts as well, though
how this might be done or under what circumstances they
should be used have not been widely considered.
Ethereum resolved the collapse of the DAO through a
so-called "fork" of the blockchain, which creates a new
chain using an updated version of the software. If enough
members of the network migrate to the new version, the
new chain takes over as the main chain, and the undesirable transactions made on the old chain are effectively
erased. Forking is similar to overriding a digital rights management system in that it allows some external power to
step in when the ordinary system can't deal with a problem, but is far more drastic in that it affects the entire
blockchain and not just the problematic smart contract.
Forks are therefore controversial and difficult to accomplish, and do not have much appeal as a routine solution
every time a contract goes awry.
We can imagine a particularly challenging scenario in
which a court finds some clause of a widely-used contract to be unenforceable according to the laws recognized by that court, as might happen if a vendor were to
include the questionable clause in the terms of use for a
popular product. Could or should the court demand that
there be a fork in the blockchain, undoing all of the
unenforceable contracts? Or could smart contract systems offer some other form of resolution, akin those
proposed for digital rights management?
Finally, the discussion so far has ignored the existence in the real world of many different jurisdictions,
each with its own version of contract law. Digital rights
management systems too largely ignored the differences
between copyright law from one jurisdiction to another:
as noted by Lessig, licenses effectively operated as
privatized law created by the organizations that issued
them, having no regard for what local law might say
about copyright. Current smart contract systems work in
similar ways, maintaining a single blockchain and algorithm across all jurisdictions.

IEEE TECHNOLOGY AND SOCIETY MAGAZINE

∕

MARCH 2020

IEEE Technology and Society Magazine - March 2020

Table of Contents for the Digital Edition of IEEE Technology and Society Magazine - March 2020