Field Actions Science Reports

The journal of field actions

Special Issue 17 | 2017
Artificial Intelligence and Robotics in the City

The Expansion of Algorithmic Governance: From

Code is Law to Law is Code
Samer Hassan and Primavera De Filippi

Electronic version
URL: [Link]
ISSN: 1867-8521

Publisher
Institut Veolia

Printed version
Date of publication: 31 December 2017
Number of pages: 88-90
ISSN: 1867-139X

Electronic reference
Samer Hassan and Primavera De Filippi, « The Expansion of Algorithmic Governance: From Code is
Law to Law is Code », Field Actions Science Reports [Online], Special Issue 17 | 2017, Online since 31
December 2017, connection on 30 April 2019. URL : [Link]

Creative Commons Attribution 3.0 License

 [Link]

THE EXPANSION
“Code is law” is a form of regulation
whereby technology is used to enforce
existing rules. With the advent of

OF ALGORITHMIC Blockchain and Machine Learning, we

are witnessing a new trend, whereby
technology is progressively taking the
GOVERNANCE: upper-hand over these rules.

From Code is Law to Law is Code

Samer Hassan Primavera De Filippi
Associate Professor, Univ. Complutense Researcher at CERSA/CNRS
de Madrid Faculty Associate at Berkman-Klein
Faculty Associate, Berkman Klein Center Center for Internet & Society
at Harvard University
INTRODUCTION
We are spending increasing amounts of our
lives interacting within platforms, whose user
base belittle that of existing nation states.
And yet, their governance is very far from the
values of democratic countries. Instead, they
are governed by software and algorithms that
regulate our interactions. As put by Lessig,
“Code is Law”, a form of regulation where
private actors may embed their values into
technological artifacts, effectively constraining
our actions. Today, code is also used by the
public sector as a regulatory mechanism. This
brings a variety benefits, mostly related to the
ability to automate the law and to enforce rules
Samer Hassan is an activist and researcher. He is Faculty and regulations a priori, i.e. before the fact. Yet,
Associate at the Berkman Klein Center for Internet & Society regulation by code also comes with important
(Harvard University) and Associate Professor at the Universidad limitations and drawbacks that might create
Complutense de Madrid (Spain). Samer received a 1.5M€ ERC new issues related to fairness and due process.
grant to build blockchain based organizations for the collaborative Blockchain technolog y comes with many
economy. He coordinated building decentralized web-tools for newfound opportunities of turning law into code.
By transposing legal or contractual provisions
collaborative communities and social movements, such as SwellRT
into a blockchain-based “smar t contract”
and Jetpad. He co-founded The Comunes Nonprofit and the Move
with a guarantee of execution, these rules
Commons webtool project. Follow Samer on Twitter: @samerP2P are automatically enforced by the underlying
Primavera De Filippi is a Permanent Researcher at the National blockchain network and will, therefore, always
Center of Scientifi c Research (CNRS) in Paris, a Faculty Associate execute as planned, regardless of the will of the
at the Berkman-Klein Center for Internet & Society at Harvard parties. This, of course, generates new problems
University, She is a member of the Global Future Council on related to the fact that no single party can affect
Blockchain Technologies at the World Economic Forum, and the execution of that code. With the widespread
founder of the Internet Governance Forum’s dynamic coalitions adoption of Machine Learning, it is possible to
circumvent some of the limitations of regulation
on Blockchain Technology (COALA). Her book, “Blockchain and
by code. ML allows for the introduction of code-
the Law,” will be published in 2018 by Harvard University Press
based rules which are inherently dynamic and
(co-authored with Aaron Wright). adaptive, replicating some of the characteristics
of traditional legal rules characterized by the
KEYWORDS flexibility and ambiguity of natural language.
• CODE-IS-LAW However, the use of ML in the contex t of
• BLOCKCHAIN regulation is not devoid of any drawback: data-
• SMART CONTRACTS driven decision making has shown implicit bias
• CRYPTOCURRENCIES that discriminate minorities, and ML-driven
• SELF-GOVERNING
laws may damage traditional principles such as
• PROTOCOLS
universality and non-discrimination.

88
 Will we succeed in making
the AI revolution work for everyone?
[Link]

To day, re g ulatio n by c o d e is p ro g re s sive ly

establishing itself as a regulatory mechanism
adopted not only by the private sector but also
by the public sector. Governments and public
administrations increasingly rely on software
algorithms and technological tools in order to
define code-base rules, which are automatically
exe c u te d (o r e n fo rc e d ) by t h e u n d e r l y i n g
technology. This is the case, for instance, of the
No Fly List in the U.S., which relies on data mining
to make predictive assessments about potential
threats to national security (Citron 2007), or the
use of computer algorithms to support judicial
decision-making and determine jail sentences or
paroles (O’Neil 2016).
Relying on technological tools and code-based
rules as a means to regulate societ y brings
about a variety benefits, mostly related to the
I. FROM CODE IS LAW TO LAW IS CODE ability to automate the law and to enforce rules
and regulations a priori, i.e. before the fact. Yet,
We are spending increasing amounts of our lives interacting within
regulation by code also come with important
platforms, whose user base belittle that of existing nation states,
drawbacks that might ultimately disrupt some of
e.g. Facebook enjoys more than 2 billion users, Youtube 1 billion, and
the basic tenets of law.
Instagram 700 million users. And yet, their governance is very far
from the values of democratic countries. Instead, they are governed On the one hand, in contrast to traditional legal
by software and algorithms that regulate our interactions and online rules, which must be appreciated by a judge and
communications through obscure rules embedded in source code, applied on a case-by-cases basis, code-based
and elaborated by a handful of private actors. rules are writ ten in the rigid and formalized
language of code, which does not benefit from the
The digital environment opens up the doors to a new form of
flexibility and ambiguity of natural language. On
regulation — by private actors — which might try to impose their own
the other hand, the architectural implementation
values by embedding them into a technological artefact. As stated
of online platforms ultimately depends on the
by Lessig (1999), “Code is Law”: code is ultimately the architecture
sp ecific choices of plat form operators and
of the Internet, and — as such — is capable of constraining an
sof t ware engineers, seeking to promote or
individual’s actions via technological means.
prevent a certain type of actions. Just like any
As more and more of our interactions are governed by software, we other technological artifact, code is not neutral,
increasingly rely on technology as a means to directly enforce rules. but inherently political: it has important societal
Indeed, as opposed to traditional legal rules, which merely stipulates implications, insofar as it might support certain
what people shall or shall not do, technical rules determine what political structures or facilitate certain actions and
people can or cannot do in the first place. This eliminate the need behaviors over others (Winner 1980).
for any third party enforcement authority to intervene after the fact,
in order to punish those who infringed the law. Software ultimately
ends up stipulating what can or cannot be done in a specifi c online
setting, more frequently than the applicable law, and possible also
much more effectively.
An emblematic example of that are digital rights management
(DRM) schemes, transposing the provisions of copyright law into “INDEED, AS OPPOSED TO TRADITIONAL
technological measures of protection (Rosenblatt, et al. , 2002), LEGAL RULES, WHICH MERELY STIPULATES
and thus restricting the usage of copyrighted works (e.g., by limiting
WHAT PEOPLE SHALL OR SHALL NOT DO,
the number of possible copies of a digital song that can be made).
The advantage of this form of regulation by code is that, instead TECHNICAL RULES DETERMINE WHAT
of relying on ex-post enforcement by third parties (i.e., courts and PEOPLE CAN OR CANNOT DO IN THE
police), rules are enforced ex-ante, making it very difficult for people FIRST PLACE. THIS ELIMINATES THE NEED
to breach them in the first place. Besides, as opposed to traditional FOR ANY THIRD PARTY ENFORCEMENT
legal rules, which are inherently fl exible and ambiguous, technical
rules are highly formalized and leave little to no room for ambiguity,
AUTHORITY TO INTERVENE AFTER THE
thereby eliminating the need for judicial arbitration. FACT, IN ORDER TO PUNISH THOSE
WHO INFRINGED THE LAW.”

89
 [Link]

Machine Learning (ML) allows software to acquire knowledge from

II. NEW CHALLENGES TO LAW IS external sources and to learn or do things that it was not explicitly
CODE: BLOCKCHAIN & MACHINE programmed to do. The availability of growing amounts of data
LEARNING (“big data”), along with the recent advances in neural networks
and data mining techniques, has led to the widespread adoption of
Machine Learning in several online platforms. With ML, it becomes
B l o c kc h a i n t e c h n o l o g y — th e te c h n o l o g y in fact possible to circumvent some of the limitations traditionally
underpinning Bitcoin — is an emergent technology associated with regulation by code. While these platforms are still
that comes with many newfound opportunities for the most part governed by a set of rigid and formalized code-
of turning law into code (De Filippi & Hassan, based rules, ML allows for the introduction of code-based rules
2016). With the advent of “smar t contracts” which are inherently dynamic and adaptive—thus replicating some
( i.e. software deployed on a blockchain-based of the characteristics of traditional legal rules characterized by the
network, like Bitcoin, and executed in a distributed fl exibility and ambiguity of natural [Link], to the extent
manner by a distributed net work of peers), that they can learn from the data they collect or receive, these
blockchain technology could revolutionize the systems can evolve constantly refining their rules to better match
way in which people coordinate themselves and the specific circumstances to which they are meant to apply.
engage in many economic transactions and social
interactions (Tapscott & Tapscott 2016). Indeed, However, the use of ML in the context of regulation is not devoid
transposing legal or contractual provisions into of any drawback. Data-driven decision making has already been
a smart contract can give rise to a new set of proven to be implicitly biased, and consequently unfair (Hardt,
code-based rules with a “guarantee of execution”. 2014). Allegedly “neutral algorithms” systematically discriminate
These rules are automatically enforced by the minority groups in their generalizations, showing results which may
underlying blockchain network and will, therefore, be catalogued, for instance, as racist or sexist (Guarino 2016).
always execute as planned, regardless of the will Moreover, if implemented into law, the dynamicity of these rules
of the parties. could undercut notions of universality ( i.e. “all are equal before
A smart contract can be implemented in such the law”) and non-discrimination. As laws are incorporated into a
a way as to make it possible for multiple parties, code-based whose rules dynamically evolves as new information is
humans or machines, to interact with each other. fed into the system, it might become difficult for people to not only
These interactions are mediated by a blockchain understand, but also question the legitimacy of the rules that are
application, controlled exclusively by set of affecting their lives on a daily basis. And as more and more of these
immutable and incorruptible rules embedded in rules can be customized and adapted to the profile of each individual
its source code. As such, smart contracts increase user, the basic principles of universality and non-discrimination that
the applicability of regulation by code, by making characterize the current legal system might be forever lost.
it possible for people to formalize contractual
agreements and economic transactions into a set
of predetermined code-based rules, which are self-
executing and self-enforcing. And to the extent
that blockchain-based networks and associated
smart contracts do not rely on any central server,
they cannot be arbitrarily shut down by any REFERENCES
single party—unless specifically provided for in
their code. This further exacerbated the problem • Citron, D. K. (2007). Technological due • Lawrence Lessig, 1999. Code and other
related to the rigidity and formality of code-based process. Wash. UL Rev., 85, 1249. laws of cyberspace. New York: Basic
Books.
regulation, in that it becomes harder for any single • De Filippi, P., & Hassan, S. (2016).
Blockchain technology as a regulatory • Bill Rosenblatt, William Trippe and
party to upgrade the code or even just to affect the technology: From code is law to law is Stephen Mooney, 2002. Digital rights
execution of that code. code. First Monday, 21(12). management: Business and technology.
New York : M&T Books.
• O’Neil, C. (2016). Weapons of math
destruction: How big data increases • Moritz Hardt, How big data is unfair:
inequality and threatens democracy. Understanding unintended sources
Crown Publishing Group (NY). of unfairness in data driven decision
“AS LAWS ARE INCORPORATED INTO A • Tapscott, D., & Tapscott, A. (2016).
making. Medium. 2014. [Link]
com/@mrtz/how-big-data-is-unfair-
CODE-BASED WHOSE RULES DYNAMICALLY Blockchain Revolution: How the 9aa544d739de
Technology Behind Bitcoin Is Changing
EVOLVES AS NEW INFORMATION IS FED INTO Money, Business, and the World. • Ben Guarino. Google faulted for racial
Penguin bias in image search results for black
THE SYSTEM, IT MIGHT BECOME DIFFICULT teenagers. (2016). Washington Post.
• V. Buterin, “Ethereum: A next-generation [Link]
FOR PEOPLE TO NOT ONLY UNDERSTAND, cryptocurrency and decentralized news/morning-mix/wp/2016/06/10/
application platform,” 2014. google-faulted-for-racial-bias-in-image-
BUT ALSO QUESTION THE LEGITIMACY OF search-results-for-black-teenagers/
• Langdon Winner, 1980. “Do artifacts
THE RULES THAT ARE AFFECTING THEIR have politics?” Daedalus, volume 109,
number 1, pp. 121–136.
LIVES ON A DAILY BASIS.”

90