Contexte gauche	Mot	Contexte droit
	AI
	AI
help_outline	Ai
	ai
	ai
	ai
	ai
	ai
	AI
The Socio-Legal Relevance of	Artificial Intelligence
Introduction:	Artificial Intelligence	and Society
recent years, the field of artificial intelligence (	AI	), in
methods for hotel bookings and	ai
assistants embodied in smart speakers.	AI	is also being applied
development, as are	AI	-assisted tools used for cancer
challenges that the use of	AI
	ai
	ai
present notion of explainability within	AI
	AI	).  [4] Thus, the focus in
on clearly defining what	AI	is according to a computer
and practically applied	AI	from a socio-legal perspective
	ai
use of, and dependence on,	AI
	ai
	artificial intelligence	that could, based in its
	ai
self-improving	artificial intelligence	—a sort of “singularity” in
super-intelligence or general	artificial intelligence	, but rather, on
contemporary, everyday versions of	artificial intelligence	in order to
adopt a broad definition of	AI	that covers a number of
with the ambition to develop	artificial intelligence	to
	ai
in the field of	artificial intelligence	, but it is also closely
component of	AI	in general, and machine learning
	ai
difference between earlier	AI	-related rules and ethical principles
now concern everyday uses of	AI	and machine learning in a
artificial intelligence. This r	ai	ses a number of
How can fairness in	AI	be understood from a socio
relate to data-dependent	AI
accountability with regards to applied	AI	be
	ai
	AI	, from a multidisciplinary and socio
society could or should supervise	AI
posed by applied	AI	. Recently, political discussions in many
Group on Artificial Intelligence (	AI	HLEG) published a draft of
guidelines for trustworthy	AI	,  [9] that resulted in a
	Artificial Intelligence	(Nationell inriktning för artificiell
to guide ethical and sustainable AI, and the use of	AI
	ai
I. Socio-Legal Challenges of Artificial Intelligence: F	ai	rness
social consequences of	artificial intelligence	, a growing understanding
	ai
	ai
	ai
	ai
reproduced or automatically strengthened by	AI	systems which often only
	ai
	ai
discussions surrounding machine learning and	AI	are often
the topic of “ethics” and	artificial intelligence	, in
on values and norms in	AI	developments, as well as—and
	ai
impact	AI	is having on us, on
	AI	governance discourse may arguably be
	ai
impact. Discussions on ethics in	AI	will
	ai
	ai
and have self-learning and	artificial intelligence
	ai
	ai
	ai
	ai
	ai
	ai
For example, a report by AI research centre	AI	Now on
	AI	is not impartial or neutral
	ai
who design, develop, and maintain	AI
of the biases embedded in	AI	systems are products of a
in the context of	AI	. Agency, it seems, is one
	ai
increasingly autonomous	AI	services that will unavoidably be
20A governance approach on	AI	expressing principles or guidelines has
with a newfound vigour. Conventional	AI
laid out principles for their	AI	development
malicious implementations of	AI	that also addresses the
that studies malicious uses of AI has called for	AI
	ai
	ai
behaviours.  [47] However,	AI	, it seems, comes with an
billions. This is a newfound	AI	-driven
code L. Lessig likely was	ai
	ai
	ai
25To demonstrate how	AI	and machine learning have become
	AI	as a social challenge, two
an	AI	context leads us to study
	ai
argued that	AI	was a tool that could
criticised for expressing a simplistic ”	AI	solutionism”—in line
	ai
	ai
EU commission’s communiqué on	artificial intelligence	.  [59] The EU
	ai
	ai
a field of studies within	AI	research that focuses on the
	ai
for the analysis of applied	AI	on markets, as
aspects of	AI	governance. A challenge, from a
	AI	Now Institute, proprietary set-ups
guided by	AI	where there is an incentive
of new technologies, here applied	AI	, the
complex	AI	processes have a direct impact
	ai
explanation of how	AI	-generated decision-making works, a
	ai
	ai
	AI	-generated processes will therefore likely
be related to how contemporary	AI	very
	ai
	ai
illegal outcomes derived from automated	AI	-driven systems.  [73
7. Explainable Artificial Intelligence (X	AI	) and Algorithm Complexity
outcomes of complex	AI	tools. Within the area of
	AI	research, a
AI) that deals with expl	ai	nability or
	ai
	ai
	ai
implications of	artificial intelligence	. Mireille Hildebrandt argues
may help us understand that	artificial intelligence	not
be a key challenge as	AI
	ai
more desirable?  [82] Data-dependent	AI	that
	ai
	ai
	ai
society. Conversely, this means that	AI	-driven analytical methods may
several reports published by the	AI
of a truly neutral stance,	AI	developers will
	ai
	AI	engineers in image analyses and
	ai
	ai
	AI	applications, particularly those that interact
Conclusions: Socio-Legal	AI	Studies
challenges of	AI	. I have drawn on socio
	ai
applied	AI	and machine learning in society
to stress the need for	AI
42The argument that designing	AI	is a normative process recognizes
implemented in	AI	developments and applications in order
address aforementioned risks. Since	AI	and machine learning, when
	ai
data-driven	artificial intelligence	. Grounds for addressing
	ai
ethical guidelines for	AI	is that the combination of
	AI	and society
trusted applications. Contemporary data-dependent	AI	should not be
challenges of	AI	transparency described above. It means
superficial layer overlying current	AI	developments in computer
expertise that can contribute to	AI
producing principles to govern applied	AI
	ai
principles for	AI	and machine learning; e.g
	ai
	ai
of the fact that automated	AI	and machine learning have the
societal impact assessments of	AI	processes in relation to consumer
challenges with applied AI is dealing with expl	ai	nability and
so-called black box applications,	AI	transparency
	ai
	ai
	ai
fairness and trust in applied	AI	through well-advised
consommation	ai
	ai
IJCAI-17 Workshop on Explainable	AI
	AI	), 2017
Machine Metaethics”,	AI	& Society, 22 (4), 2008, p
	AI	HLEG, “Draft Ethics Guidelines for
Trustworthy	AI	,” 18 December
Id., Ethics Guidelines for Trustworthy	AI	, Brussels: The European
social and legal consequences of	AI	, see Stefan Larsson
Anna Felländeret al., Sustainable	AI	: An Inventory
Related to Artificial Intelligence, Stockholm: AI Sust	ai	nability
transparency” with regards to	AI	, see Stefan Larsson and Fredrik
Heintz, “	AI	Transparency”, Internet Policy Review, 2019
	ai	-Wei Chang. “Men also like
	ai
Beyond State v Loomis:	Artificial Intelligence	, Government
	ai
	ai
	ai
	ai
	ai
Stefan Larsson and Fredrik Heintz, “	AI	Transparency”, op. cit
Crawford, Roel Dobb et al.,	AI	Now Report
2018, New York:	AI	Now Institute, 2018
	ai	, “The Social, Political, Economic, and
	ai	’i
Crawford, AI Now 2017 Report,	AI	Now Institute at New York
Sundar Pichai, “	AI	at Google: Our Principles”, Google
topics/ai/	ai	-principles
Reportedly Leaving Project Maven Military	AI
ery-	ai	-contract-expire> (last visited 10
al., The Malicious Use of	Artificial Intelligence
	ai
al., The Malicious Use of	Artificial Intelligence
	ai
	ai
	Artificial Intelligence	in Governing Communication Online”, Social
Fully Solve Its Problems With	AI	”, by Davey
	ai
the Committee of the Regions,	Artificial Intelligence	for Europe
Persuasive Technology on Internet Platforms”,	AI	Now
	ai
of	AI	Under the Law: The Role
	ai
	ai
	ai
Tim Miller, “Explanation in	Artificial Intelligence	: Insights from
the Social Sciences”,	Artificial Intelligence	, 267, 2019, p. 1-38
and Londa Schiebinger, “	AI	Can Be Sexist and Racist
	ai
Crawford, Roel Dobb et al.,	AI	Now
in line with for example	AI	HLEG’s Ethics guidelines for
trustworthy	AI	(2019); the IEEE’s Ethically
	AI
Ethical Framework for a Good	AI	Society
Franç	ai
	ai
	ai
	ai
d’explainable AI (X	AI	) dans la recherche en
	ai
	ai
	ai
Explainable	AI	et transparence des algorithmes
	ai
applied artificial intelligence (	AI	) and machine learning. The purpose
normative challenges posed by applied	AI	. To do so, the article
may complement notions of explainable AI (X	AI	) within
	AI	-research undertaken by computer scientists
	ai
	AI	research, development, and governance
Applied	artificial intelligence
Explainable	AI	and algorithmic transparency
1. Introduction:	Artificial Intelligence	and Society
I. Socio-Legal Challenges of Artificial Intelligence: F	ai	rness
	ai
7. 7. Explainable Artificial Intelligence (X	AI	) and
5. Conclusions: Socio-Legal	AI	Studies
	ai
	ai
Citer cet article Franç	ai
	Artificial Intelligence	», Droit et société, 2019
	ai
	Artificial Intelligence	», Droit et société, vol
	Artificial Intelligence	. Droit et société, 103
	ai
	ai
	ai
	ai
	AI
	AI
	ai
	ai
	ai
Cairn Sciences (Franç	ai
	Ai
	ai
	ai
	ai
	ai
	ai
	ai
	ai