See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/347344369

Artificial intelligence and social responsibility: the case of the artificial

intelligence strategies in the United States, Russia, and China

Article in Kybernetes · December 2020

DOI: 10.1108/K-01-2020-0060

CITATIONS READS

31 1,554

2 authors:

Anton Saveliev Denis Zhurenkov

7 PUBLICATIONS 39 CITATIONS
Federal state unitary enterprise "All-Russia scientific and research institute "Center"
20 PUBLICATIONS 48 CITATIONS
SEE PROFILE
SEE PROFILE

All content following this page was uploaded by Denis Zhurenkov on 03 September 2021.

The user has requested enhancement of the downloaded file.

 The current issue and full text archive of this journal is available on Emerald Insight at:
https://www.emerald.com/insight/0368-492X.htm

Case of the
Artificial intelligence and social artificial
responsibility: the case of the intelligence
strategies
artificial intelligence strategies in
the United States, Russia, and China
Anton Saveliev and Denis Zhurenkov Received 31 January 2020
Revised 3 September 2020
Philosophy of science and technology, Institute of Philosophy, Russian Academy of Sciences, 29 September 2020
Moscow, Russian Federation Accepted 19 October 2020

Abstract
Purpose – The purpose of this paper is to review and analyze how the development and utilization of
artificial intelligence (AI) technologies for social responsibility are defined in the national AI strategies of the
USA, Russia and China.
Design/methodology/approach – The notion of responsibility concerning AI is currently not legally
defined by any country in the world. The authors of this research are going to use the methodology, based on
Luciano Floridi’s Unified framework of five principles for AI in society, to determine how social responsibility
is implemented in the AI strategies of the USA, Russia and China.
Findings – All three strategies for the development of AI in the USA, Russia and China, as evaluated in the
paper, contain some or other components aimed at achieving public responsibility and responsible use of AI.
The Unified framework of five principles for AI in society, developed by L. Floridi, can be used as a viable
assessment tool to determine at least in general terms how social responsibility is implied and implemented in
national strategic documents in the field of AI. However, authors of the paper call for further development in
the field of mutually recognizable ethical models for socially beneficial AI.
Practical implications – This study allows us to better understand the linkages, overlaps and
differences between modern philosophy of information, AI-ethics, social responsibility and government
regulation. The analysis provided in this paper can serve as a basic blueprint for future attempts to define
how social responsibility is understood and implied by government decision-makers.
Originality/value – The analysis provided in the paper, however general and empirical it may be, is a
first-time example of how the Unified framework of five principles for AI in society can be applied as an
assessment tool to determine social responsibility in AI-related official documents.

Keywords Artificial intelligence, Social responsibility, National strategies, Ethics, Social responsibility,
National strategies
Paper type Research paper

1. Introduction
Artificial intelligence (AI) is often understood as a double-edged sword of modern
science – a technology whose destructive potential is virtually unlimited without
proper guidance from its creator. For example, such prominent scientist and public
intellectuals as Stephen Hawking and Martin Rees, as well as innovations
entrepreneur Elon Musk and AI-researcher Stuart Russel have quite eloquently

The Authors would like to thank Dr. Igor Perko, Dr. Francesco Caputo and Dr. Teodora Ivanusa for Kybernetes
their helpful advice on various technical and philosophical issues, examined in this paper. The © Emerald Publishing Limited
0368-492X
Authors, however, bear full responsibility for the Paper. DOI 10.1108/K-01-2020-0060
K spoken about the disruptive power the AI possesses, mentioning above all the risks of
total annihilation of mankind, should strong AI-technology go rogue, or fall into
wrong hands. Public perception of AI-related risks spreads even further, calling for
collective action.
In 2009, leading AI-researchers (Horvitz and Selman, 2009) voiced their rising concern
during the Asilomar conference, which eventually led to the signing of an open letter (Future
of Life Institute, 2015), and the creation of Asilomar AI Principles (Future of Life Institute,
2017) – an assortment of 23 guidelines, outlining AI developmental issues and ethics for the
development of beneficial AI. On the other hand, potential dangers of weaponized AI-
applications were openly discussed at a leading AI conference, the Association for the
Advancement of Artificial Intelligence (AAAI) 2015, and at a workshop on AI and ethics of
the same conference. The international community (UNESCO, 2018; Floridi and Cowls, 2019)
also shared a certain mistrust of AI, but in general agreed that it is difficult to develop and
formalize any clear guidelines to avoid major risks: the AI technology is too complex.
All this indicates that the AI is no longer considered a purely technological phenomenon,
but rather a social one. While the possible impact of AI-related technologies on modern
society is yet unknown, researchers, enthusiasts and political leaders are still struggling to
define a framework, which will ensure the utilization of AI for the benefit of social
responsibility. With the growing interest in AI, more and more countries are creating their
national strategies to develop and harness this technological marvel of the 21st century.
These pioneering documents often seem vague and uncertain, but still offer considerable
insight on how AI will affect people’s lives, safety and well-being, should governments in
technologically advanced countries fully embrace this technology. This fact alone is worthy
of thoughtful review and analysis, to find out, how these strategies define (if define at all) the
use of AI for social responsibility. For this purpose, the authors have chosen official
strategic documents in the field of AI development and implementation of the three
technologically advanced countries – the USA, Russia and China. These documents will be
examined through the methodology, based on the Unified framework of five principles for AI
in society. Italian philosopher and researcher Luciano Floridi put forward the idea of basic
ethical principles of AI, using his concept of Philosophy of Information. These principles
now form the basis of the EU Ethics guidelines for trustworthy AI. In this paper, we are
going to assess, how the Unified framework of five principles for AI in society can be
applied for a better understanding of social responsibility’s role in AI-related official
documents. The choice of Luciano Floridi’s theoretical contribution as a basis for this
research methodology is justified by the fact that it allows us to consider AI as an
autonomous agent, and thus enables us, at least as a part of the theoretical assumption, to
use the concept of “responsibility” with respect to AI.

1.1 Conceptual background

Currently, the ethical and social impact of AI can be roughly divided into two distinct areas
of study – the human perspective and the AI perspective. This distinction is not dictated by
the primordial nature of AI a possibly alienated (in logic and behavior) entity, but rather by
our incomplete understanding of its role. The human perspective is of paramount
importance for us and will be briefly reviewed in this section.
John P. Sullins in his works argues that a state of moral agency can be ascribed to AI,
especially if this AI fulfills a social role. In this case, it must possess a duty of care, which is
only possible if it is a moral agent (Sullins, 2006). Sean O’Heigeartaigh defines the perils of
moral outsourcing in connection to AI: human biases are numerous but they are relatively
easy to grasp and to correct. A machine error, on the contrary, can lead to unthinkable
catastrophic consequences (O’Heigeartaigh, 2013). Even more so, certain researchers believe, Case of the
that moral outsourcing cannot be chastised even by the means of strict algorithmic artificial
accountability (Pavaloiu and Kose, 2017). In this regard, P. Boddington points out that virtue
ethics and Kantian morality do not allow for the outsourcing of moral judgments, since
intelligence
another machine, however, sophisticated, would be unable to do the right things for the right strategies
reasons and in the right manner (Boddington, 2017). On the other hand, some scientists
advocate the delegation of morality to computer systems. Jos de Mul argues that computer
systems can act as important moderators of human agency, enhancing our moral autonomy
(de Mul, 2010). Jeffrey K. Gurney explores AI ethics through the framework of a crash-
optimization algorithm, in which the algorithm author allows the autonomous machine to
decide who and what to hit under certain conditions, dictated by classical moral dilemmas
(Shopping Cart Problem, Motorcycle Problem, The Car Problem, etc.) (Gurney, 2016).
Issues of socially responsible application of AI are addressed by J. Borenstein and Y.
Pearson. Using a capabilities approach analysis, they argue, that robots can indeed enhance
the caregiving process and empower caretakers with more freedom in their life (Borenstein
and Pearson, 2010). Sharkey (2008), however, points out the dangers of this approach:
caretakers can harbor possibly unhealthy attachment to inanimate entities: i.e. elderly, left in
the exclusive care of machines, would be deprived of the human contact. J. Kofas proclaims
that AI will significantly impact the sense of identity and community in society, by
undermining the community culture and widening inequality in society. He argues, that AI
and robotics will enhance living standards, but the most valued applications of these
technologies will be available exclusively to wealthy classes (Kofas, 2017).
The ethical and social shortcomings of algorithms (which are the main pillars of AI) are
addressed by Friedler et al. who examine the meaning of a “fair algorithm” on the basis of
philosophy and computer science. Researchers provide a mathematical definition of
“fairness”, and demonstrate that fairness in output depends on the interactions between the
construct space, observed space and the decision space of the algorithm (Friedler et al.,
2016). Anderson and Sharrock point out differences between algorithms and social actors:
the former are bound by mathematical instructions, the latter can exercise discretion. Yet
they believe that algorithms can be relied on to make satisfactory ethical decisions
(Anderson and Sherrock, 2013).
Speaking about the general design perspectives of AI ethics, Pavaloiu and Kose remind
us about data interpretation bias, found in the top-down approach: e.g. the AI might
indirectly infer if an individual is depressed based on social media data and this can affect
future employment of this individual (Pavaloiu and Köse, 2017). Allen et.al state that the top-
down approach permits the architect to tailor the capacity of the system in design. At the
same time, there can be a possible conflict between the pre-encoded rules and the constant
pressure to predict and compute outcomes for every action. To prevent this conflict we are
going to need a universal standard for AI ethics since computers are still limited in their
computational capacity and thus will struggle to find the right outcome for every complex
ethical dilemma. Meanwhile, the bottom-up approach can be based on the rewards and
punishments system, in which AI can be compared to a child brought up in a rough
neighborhood. In conclusion, researchers call for a hybrid model which, while framing broad
governing rules, would provide scope for learning through experience (Allen et al., 2005).
There are certain developments in specialized applicable standards, such as for the
protection of personal data (Heisenberg, 2005; Albrecht, 2016) or in robotics (Schlenoff, 2015;
Noorman and Johnson, March 2014), as well as in the research on the legal personality of AI
(Gadzhiev, 2018; Arkhipov and Naumov, 2017; Morkhat, 2018; Cerka et al., 2017; Morkhat,
2017). In this regard, we would like to point out the works of the Russian Research Center for
K the Problems of Regulation of Robotics and AI (Arkhipov and Naumov, 2017), which is
drafting legislation in the field of robotics and AI (http://robopravo.ru).
While the above cited studies do indicate a growing interest on behalf of an academic
community, one is forced to agree that a general and universally recognized legal or ethical
system of AI-related risks control does not exist (Darnault, et al. August 2019). Nonetheless,
this issue does not stop technologically advanced nations from creating and formalizing
their own AI-development and utilization strategies, aimed to better adapt the AI
technologies for the benefits of security and economic prosperity. Despite being state-
mandated, these strategies are not restrictive, nor clearly defined, focusing on all AI-related
issues and matters at once. In the absence of any specific AI-control laws of national or
international origin, even these vague documents are still a great source of insight. AI is no
longer just a promising technological gimmick, instead, it is perceived as a social, political,
and history-changing phenomenon, easily comparable to the atomic bomb, the infamous
brainchild of 20th century physics.
In addition, just like with the nuclear fissure, society is both scared of its power and eager to
harness it in. This paper explores how responsibility to society, as well as any means to
facilitate social responsibility, is defined (if defined at all) in national AI development strategies
in three technologically advanced countries – the USA, Russia and China. To achieve this task,
American, Russian and Chinese strategies will be reviewed within the combined framework of
existing voluntary applicable standards in the field of AI and the theoretical basis of modern
philosophy of information. As authors, we must state, that the choice of these three countries
for our analysis is not justified by their ethical contribution to the seemingly different domains
of social responsibility and AI. On the contrary, our decision was dictated by the interest in the
development of AI technologies in these countries on the part of government authorities and
major market players, which found its reflection in official documents.
USA and China are currently leading the AI race in terms of technological advance and
market application, while Russia is but following in their steps, trying to emulate successful
approaches set by both countries. While not possessing access to autonomous robotic
systems, Russia lags in R&D and spending on AI technologies as well. The Russian
government’s future investment in AI research is unknown, Brookings Institution report
estimates that the country spends approximately $12.5m a year on AI research, putting it far
behind China’s plan to invest $150bn through 2030. The US Department of Defense alone
spends $7.4bn annually on unclassified research and development on AI and related fields
(Brookings Institution, 2018). Nonetheless, despite these disparities, Russia is still capable to
become a local leader in applied AI technologies due to relatively advanced IT-infrastructure
and great fundamental scientific potential.

2. Theory and methodology

AI is an area of information technology that deals with the development of intelligent
systems, i.e. systems possessing capabilities that are traditionally associated with the
human mind – language understanding, learning, ability to argue, solve problems, etc.
Moreover, the terminology of AI systems also includes the following concepts:
 Narrow AI, which has been available for decades and is still widely used today.
Narrow AI systems are complex software programs that can execute discrete
‘intelligent’ tasks such as recognizing objects or people from images, translating
language, or playing games. Narrow AI systems can execute complex calculations,
but they are limited by the boundaries of their task, operating environment and
specified programming (SIPRI, 2019).
  Autonomous AI, which provides the ability of the system to function for a long time Case of the
without the participation of an operator. artificial
 Adaptive AI, which implies the ability of the system to adapt to new conditions, intelligence
acquiring knowledge that is not embedded in the creation.
strategies
 Artificial General Intelligence (AGI), which is characterized by such high
adaptability that the system possessing it can be used in a variety of activities with
appropriate training, both independent and directed (with the help of an instructor).
 Strong AI (Strong AI), Human-Level AI (Human-Level AI) with a level of
adaptability comparable to the human level.
 Super-human AI (Super-human AI), which implies even higher adaptability and
speed of learning.

Most of the aforementioned paradigms still satisfy the definition, proposed by John
McCarthy, Marvin Minsky, Nathaniel Rochester and Claude Shannon in seminal Proposal
for the Dartmouth Summer Research Project on AI (McCarthy et al., 2006 [1955]):
For the present purpose the artificial intelligence problem is taken to be that of making a machine
behave in ways that would be called intelligent if a human were so behaving.
However, this classic definition is viewed by some researchers as counterfactual. Italian
philosopher Luciano Floridi, whose theoretical findings are extensively used in this paper,
states that:
Just because a dishwasher cleans the dishes as well as (or even better than) I do does not mean
that it cleans them like I do, or needs any intelligence to achieve its task (Floridi and Cowls, 2019).
Instead, Floridi provides his broad definition, which is nonetheless still compatible with the
classical one, and conceptualizes AI as:
[. . .] a growing resource of interactive, autonomous, and often self-learning agency [. . .], that can
deal with tasks that would otherwise require human intelligence and intervention to be performed
successfully. In short, AI is defined based on engineered outcomes and actions and so, in what
follows, we shall treat AI as a reservoir of smart agency on tap[. . .] (Floridi, 2019).
This definition is highly relevant for this research, as it enables the use of “responsibility”
concerning AI applications. Stanford Encyclopedia of Philosophy declares, that “In very
general terms, an agent is a being with the capacity to act, and agency denotes the exercise
or manifestation of this capacity” (Stanford Encyclopedia of Philosophy, 2019). Therefore, if
we embrace Floridi’s definition of AI, we are allowed to empower AI and robotic systems with
a limited ability to act autonomously and even independently at least in the imagination of a
scientist exploring the hypothetical reality. This assumption is a cornerstone of this research,
which argues that responsibility (social or other) cannot be achieved without the ability to
act without autonomy or independence. It is true that among the society of homo sapiens
species there are individuals capable of independent actions, but wholly or partially devoid
of responsibility for them, yet the legislative systems, currently employed in the world, do
not possess the status of a responsible individual incapable of independent action.
Social responsibility is an ethical theory in which individuals are accountable for fulfilling
their civic duty, and the actions of an individual must benefit the whole of society. ISO 26000,
the international standard for social responsibility declares, that there must be a balance
between economic growth and the welfare of society and the environment. If this equilibrium is
maintained, then social responsibility is accomplished, and it will guide us towards greater/
social good. It is a rather simple consequentialist approach, which leads us to a complex
K dilemma. Consequentialism-based ethical constructs mandate us to increase the probability of
good consequences, but they fail to define what these potential consequences are. In other
words, the exact meaning of “socially responsible” and “good” depends solely on what the
agent in question considers as such. However, such an approach could cause dangerous, if not
life-threatening, problems when applied to AI. What if an AI system makes a decision, which is
not universally recognized as “good” in nature by society? Should a self-driving car prioritize
potentially expensive damage to public property and environment over the life of a single
human in a dangerous situation? And what about “augmented intelligence” systems which
involve an extensive amount of networked human-to-machine and machine-to-machine
communications? What happens if such a system gets tampered with to produce potentially
harmful effects?
These questions cannot be answered until we define the ontological and ethical status of
objects (or subjects?), which possess or operate AI. L. Floridi’s Philosophy of information
defines information object as any and every part of animate and inanimate nature,
everything that existed, exists or will exist (may exist), everything that somehow or other
participates in communication interactions, ensures the emergence, transfer, or preservation
of knowledge, that is, ensures the existence of the infosphere. The modern world is filled
with information objects (robots, programs, engineering artifacts, etc.), which cannot be
called alive, but they have high value and importance in the system of information
exchange. In this regard, L. Floridi proclaims the highest value of information and
information objects as the basis for new information ethics. Everything that is information
and contributes to its multiplication has value: artifacts, distant stars, past and future
generations, ancient civilizations, and modern computer programs. In this case, the absolute
evil is any destruction of information and impoverishment of the infosphere (entropy, both in
L. Floridi’s and C. Shannon’s terminology) (Floridi, 2002; Shannon, 1948).
In this regard, L. Floridi proclaims that information and information objects, inhabiting
the infosphere, are to be considered as the ultimate value of macro-ethics (Floridi, 2008). Evil
is everything that causes or promotes entropy, good is everything that stimulates the
growth of the infosphere. A moral act is evaluated in terms of the well-being of the
infosphere, and it can be performed even by an inanimate information agent. For example, a
computer program that protects data and makes calculations in this sphere does a positive
thing. Destruction of organized and coherent data and destruction of the entire information
system is considered as a negative action. This philosophical view of the ontological and
ethical status of information objects allows us to overcome the anthropocentric bias, without
renouncing our human and biological origins, and to establish a valid connection with a
great heritage of known and proven ethical systems. For this purpose, L. Floridi creates a
concept of “inforg” an informationally embodied organism, entity made up of information,
that exists in the infosphere. These organisms are made up of information, thus they differ
from natural agents with whom they coexist. This concept is quite similar to Norbert
Wiener’s view of organisms as entities defined by patterns of persisting “Shannon
information” – information placed into the physical realm to be manipulated by the laws of
nature and science (Wiener, 1961). This means that inforgs consist of matter, energy, and
Shannon information. The concept of inforg in relation to the information is not only
philosophical and cybernetical but biological as well. Recent experiments show, that it is
possible to use DNA as data storage, meaning that binary helix can be encoded to hold
binary information (Goldman et al., 2013), which in its own turn reinforces the concept of
living organisms as “persisting patterns of Shannon information encoded within an ever-
changing flux of matter-energy” (Bynum, 2010). Speaking of persisting patterns, one must
keep in mind that the Shannon information within an inforg contains the identity of said
organism. Human identity is not defined by matter, but rather by encoded patterns of Case of the
Shannon information within the body. Our bodies are adaptive organisms, forever changing artificial
in response to their environment, but our identity persists through time and (to certain
extent) governs this constant adaption. Thus, we are able to manipulate Shannon
intelligence
information in the metaphysical realm to influence (and govern) the physical one. This strategies
alludes that information ethics is crucial in a viable systems approach to the solution of the
growing number of challenges brought forth by the digital age and information revolution.
However, what are the key principles of information ethics in relation to AI and society?
If we are already living in the Universe, defined by information on both physical and
metaphysical levels, does it mean that there is a possible underlying connection between
these planes of our existence? In his joint article with Josh Cowls “A Unified Framework of
Five Principles for AI in Society” L. Floridi analyzed six high-profile initiatives established
in the interest of ethical and socially beneficial AI, including:
 The Asilomar AI Principles, developed under the auspices of the Future of Life
Institute in 2017 (Future of Life Institute, 2017).
 The Montreal Declaration for Responsible AI developed under the auspices of the
University of Montreal in 2017 (Montreal University, 2017).
 The General Principles offered in the second version of Ethically Aligned Design: A
Vision for Prioritizing Human Well-being with Autonomous and Intelligent
Systems. This document is a joined crowdsourced effort of 250 global scientific and
academic leaders under the patronage of the Institute of Electrical and Electronics
Engineers (IEEE) (IEEE, 2017).
 The Ethical Principles offered in the Statement on AI, Robotics and ‘Autonomous’
Systems, published by the European Commission’s European Group on Ethics in
Science and New Technologies, in March 2018 (European Commission, 2018).
 The ‘five overarching principles for an AI code’ offered in the UK House of Lords AI
Committee’s report (House of Lords, 2018).
 The Tenets of the Partnership on AI, a multi-stakeholder organization consisting of
academics, researchers, civil society organizations, companies building and utilizing
AI technology, and other groups (Partnership on AI, 2018).

Floridi and Cowls have reviewed all these documents and found that together they yield 47
basic principles of how AI can be used in a socially beneficial way. According to the concept,
embraced by the aforementioned researchers, all these principles have a high degree of
cohesion with the four core principles commonly used in bioethics: beneficence, non-
maleficence, autonomy, and justice (Floridi and Cowls, 2019; Beauchamp, 2012). According to
Floridi, bioethics is the one that most closely resembles information ethics in dealing
ecologically with new forms of agents, patients and environments (Floridi, 2013). However,
Floridi and Cowls argue that a new principle is needed in addition: the principle explicability,
which is “understood as incorporating both intelligibility (for non-experts, e.g. patients or
business customers, and experts, e.g. product designers or engineers) and accountability”
(Floridi and Cowls, 2019). Thus, the complete Unified framework of five principles for AI in
society (hereinafter referred to as framework), according to Floridi, would look like this:
(1) Beneficence: promoting well-being, preserving dignity, and sustaining the planet.
This principle, found to some extent in all six initiatives, firmly underlines the
central importance of promoting the well-being of people and the planet with AI
K technologies, continued prospering for mankind and the preservation of a socially-
responsible environment for future generations;
(2) Non-maleficence: privacy, security, and ‘capability caution’. This principle cautions
against various negative consequences of overusing or misusing AI technologies,
especially when dealing with personal privacy and militarized AI applications.
However it is still unclear what the exact nature of the possible AI-maleficence would
look like: are we dealing with the people developing AI, or the technology itself.
(3) Autonomy: the power to decide (to decide). “When we adopt AI and its smart
agency, we willingly cede some of our decision-making power to technological
artifacts”, states Floridi. This important principle speaks for a balance between the
decision-making power people retain for themselves and the power they delegate
to artificial agents. Humans should retain the power to decide which decisions to
take: exercising the freedom to choose where necessary, and ceding it in cases
where overriding reasons, argues Floridi.
(4) Justice: promoting prosperity, preserving solidarity, avoiding unfairness. While
justice may seem as a rather broad concept, all AI enthusiasts and thinkers agree
that elimination of unfair discrimination and the need for shared prosperity should
stand in the center of AI use.
(5) Explicability: enabling other principles through intelligibility and accountability.
Simply put, this principle must answer the question: is humanity the patient,
receiving the ‘treatment’ of AI, the doctor prescribing it, or possibly both. So the
explicability principle really should incorporate both the meaning of how AI works
and the meaning of who is responsible for the way it works. Floridi argues, that
this principle complements the others: for AI to be beneficent and non-maleficent,
we must be able to understand the responsibility or harm it is doing to society, and
in which ways (Floridi and Cowls, 2019).

At first glance, the framework developed by Floridi is sufficient to analyze any national
strategy in the field of AI: such documents are full of broad and declarative statements of
general “do responsibly” nature. However, the question is to what extent these ethical
principles (borrowed from bioethics) are appropriate for assessing the social responsibility
in AI utilization. Sure, Floridi’s framework was adopted by the EU and is now an integral
part of AI Ethics Guidelines of the European Commission’s High-Level Expert Group on
Artificial Intelligence (AI HLEG), but maybe there is a deeper connection between the
seemingly different worlds of silicon and carbon, and not just our desire to use old laws for
new “trespassing”? Our modern reality is inexorably hybrid and not only in terms of various
“interreality systems” we use constantly – from advanced heads-up displays in planes and
even cars to QR-code scanners in smartphones. Virtual tools, coupled with their real-world
counterparts, inevitably lose their “virtuality” if only in the eyes of the common user, who
becomes increasingly reliant (if not dependent) on them, just like he relies on caffeine and
painkillers in everyday life, and probably even more so. In this respect, we can hypothesize
that AI is no longer simply a tool for us to wield, nor is it a lifestyle application, but a
completely new dimension to the informatisized reality we live in. And as much as we would
like to present this process as something completely new to society, we might just discover
instead that it is not so novel in fact. In the modern world, there is a saying: “The human
being is the sum of social relations” that can be paraphrased as “The human being is the
sum of information and its transmission channels”. Human lifestyle as biosocial creatures is
very much hybrid in nature: we strive to create our own metaphysical reality through
culture and thinking and we cannot fully if ever at all, leave the inherent reality of nature. As Case of the
a result, what we get is an entanglement, or co-production of nature/culture, and not simply artificial
modern social and cultural phenomena cast in biological metaphors (Rabinow, 1996). And
the more we embed AI into our society and culture, the more it becomes entangled into the
intelligence
surrounding nature as well as the nature of humanity as species and information agents. strategies
Ultimately, we can agree with Floridi and propose the working hypothesis that bioethics
and the ethics of information and AI can indeed overlap and complement each other. Thus,
the Unified framework of five principles for AI in society can be temporarily adopted as a
potentially viable assessment tool for how social responsibility is implied and implemented in
AI-development strategies. To make this task a little bit easier, the authors would like to
propose the following visualization of the chosen assessment tool (Figure 1).
The diagram presented above is a visualization of the analysis of AI development
strategies in this article. For lack of a more precise methodology, we will empirically assess
the positions of each of the considered strategies with the Unified framework, where the blue
diagram corresponds to each case of the exact/relative resemblance of a particular principle
(beneficence, non-maleficence, autonomy, and justice), and the orange diagram – to how
explicit each of the principles is represented (e.g. how well it is supported, what are the
means to implement it, etc.). The scale of the blue and orange diagrams is measured
according to the following grade system (Table 1).
The final verdict will be based on a simple semantic analysis, where the framework grade
units, chosen above, may or may not encounter their semantic representation in the analyzed
text. The authors of this paper do not aim to determine the exact frequency of these encounters.
Instead, they will try to examine a basic semantic correlation between the Unified framework
provisions and those of the analyzed documents. This methodological perspective may appear
to be too vague, but the authors fear that any other more specialized approach could be too
narrow for the documents which appear to be vague in the letter as well as in spirit.

3. National artificial intelligence strategies and social responsibility

3.1 The USA
The USA could be considered as an active leader in terms of AI technologies. It is worth to
mention, that the majority of global companies in the AI industry are concentrated there – over
2,000 US-based companies are leading the way in terms of AI start-ups – 1,393 (over 54% of the
global market) compared to 389 in China (about 15%). Nevertheless, despite this unquestioned
market leadership, the country still struggles to establish an active legal regulation in the field
of AI. The closest thing the USA has to the national strategy is the Executive Order on AI,

Beneficence
2
1.5
1
0.5
Non-
Jusce 0
maleficence

Figure 1.
Assessment tool for
the framework of five
principles for AI in
Autonomy society
Direct representaon Explicable representaon
K Basic principle representation (Blue Explicability of the principles (Orange
Grade diagram) diagram)

0 The principle is not mentioned

No supporting means mentioned and no
further representation
1 The principle is mentioned at least once in
the document Supporting means are mentioned, but are
purely declarative
2 The principle is detailed in at least one of
the document’s strategic goals Supporting means are detailed and are
Table 1.
backed by legal initiatives
Grade system for the 3 The principle is represented in numerous
framework of five strategic goals of the document Supporting means are well detailed, are
principles for AI in backed by legal initiatives and have
society assigned government-assisted programs

signed by President D. Trump on February 11, 2019 (Executive Office of the President, 2019).
The document announced the American AI Initiative – the USA’ national strategy on AI. This
strategy is a concerted effort to promote and protect US AI technology and innovation. The
Initiative implements a whole-of-government strategy in collaboration and engagement with
the private sector, academia, the public and like-minded international partners. It directs the
Federal government to pursue five pillars for advancing AI:
(1) promote sustained AI R&D investment;
(2) unleash Federal AI resources;
(3) remove barriers to AI innovation;
(4) empower the American worker with AI-focused education, and training
opportunities; and
(5) promote an international environment that is supportive of American AI
innovation and its responsible use.

The USA is also actively leveraging AI to help the Federal government work smarter in its
processes and services.
Executive Order on AI is further detailed and supported by the National AI R&D
Strategic Plan of 2019, which establishes the following eight strategic priorities:
(1) make long-term investments in AI research. Prioritize investments in the next
generation of AI that will drive discovery and insight and enable the USA to
remain a world leader in AI;
(2) develop effective methods for human-AI collaboration. Increase understanding of
how to create AI systems that effectively complement and augment human
capabilities;
(3) understand and address the ethical, legal, and societal implications of AI. Research
AI systems that incorporate ethical, legal, and societal concerns through technical
mechanisms;
 (4) ensure the safety and security of AI systems. Advance knowledge of how to design Case of the
AI systems that are reliable, dependable, safe, and trustworthy; artificial
(5) develop shared public datasets and environments for AI training and testing. intelligence
Develop and enable access to high-quality datasets and environments, as well as to strategies
testing and training resources;
(6) measure and evaluate AI technologies through standards and benchmarks.
Develop a broad spectrum of evaluative techniques for AI, including technical
standards and benchmarks;
(7) better understand the national AI R&D workforce needs. Improve opportunities
for R&D workforce development to strategically foster an AI-ready workforce; and
(8) expand public-private partnerships to accelerate advances in AI. Promote
opportunities for sustained investment in AI R&D and for transitioning advances
into practical capabilities, in collaboration with academia, industry, international
partners, and other non-Federal entities.

Each of the aforementioned priorities is a complete strategy in itself, with corresponding

R&D programs and goals (United States Government, 2019).
A unified framework of five principles is rather easily applicable to the US AI strategy:
 Beneficence (1) is perhaps the most represented principle here, as all eight strategic
priorities are mentioned to be potentially beneficial to “nearly all aspects of society,
including the economy, healthcare, security, the law, transportation, even
technology itself.” The benefits to the society are described to be numerous, with
enhanced quality-of-life through the use of AI in disaster recovery and medical
diagnostics, new workplaces on th0065 market, etc. On the basic scale, the
beneficence principle in the US AI strategy can be assigned grade “3”. However, on
the explicability scale, this principle is not very well represented at all. There are no
legal initiatives and mentioned priorities do not define specific research agendas for
Federal agency investments. Nonetheless, they still provide an expectation for the
overall portfolio for US Federal AI R&D investments, thus grade “1.5” can be
assigned (Figure 2).
 Non-maleficence (2) is indirectly mentioned in goal 3. Understand and Address the
Ethical, Legal, and Societal Implications of AI, which calls for ethical and secure AI
deployment to prevent all possible harm to humans. Besides, non-maleficence is

Beneficence
3

1
Non-
Jusce 0
maleficence

Figure 2.
The framework of the
Autonomy
US AI strategy
principles
Direct representaon Explicable representaon
K indirectly implied in goal 4. Ensure the Safety and Security of AI Systems. It is
stated, that AI should be “safe by design”, where safety is not the only concern of
system designers, but also is considered throughout the AI system lifecycle. On the
basic scale we can assign grade “2.5”, but on the explicability scale only grade “1” is
possible since all supporting means are declarative.
 Autonomy (3) principle is represented in strategic goal 2. Develop Effective Methods
for Human-AI Collaboration, where it is stated that the deployment of AI systems
should be considered “as a design option for operators who wish to decide whether
they want to use them or not”. Grade 1 can be safely assigned on the basic scale, but
the total lack of supporting means and further legal representation means only
grade “0” on the explicability scale.
 Justice (4) is well represented in the US strategy both in the meaning of just and fair
AI use (extensively detailed in goal 3. Understand and Address the Ethical, Legal,
and Societal Implications of AI) and “social justice” as well (goal 7. Better
Understand the National AI R&D Workforce Needs). The latter calls for the need to
support American researchers in the field of AI, as well as undergraduate students.
On the basic scale grade “3” is assigned, but only grade “1.5” on the explicability
scale: numerous supporting means are declared, albeit without any legal and
financial backing.

3.2 Russia
In the past 10 years in Russia, some 1400 AI scientific projects were carried out. Yet most of
them (1200) were non-profit, as the private sector did not show much interest in the
development and use of AI. Hence, AI was mostly driven by state and state-owned
businesses. In the last 10 years, Russia allocated about $343m for R&D in AI. In comparison,
the US state budget allocates about $200m annually for research in AI.
The Russian National AI Development Strategy was published on the 10 of October
2019. Until this date, the country lacked clear strategic documents, outlining AI
development, use, and priorities. Russian Strategy mostly focuses on civilian AI
applications (academia, science, healthcare, infrastructure, etc.), where new technologies can
provide fast if not an instant benefit to the public sector.
Russian AI Development Strategy can be divided into six blocks, each corresponding to
the certain national priority field:
(1) Legal and ethical framework: Establish clear legal foundations for creating and
running industrial and academic centers for interdisciplinary applied research;
determine the distribution of responsibility between owners, developers, and
suppliers of data for AI use; clarify the regulation of the circulation of results of
intellectual activity using AI; create national standards for testing, certifying, and
confirming compliance for AI systems.
(2) Scientific and academic community: Create a new methodology for calculating a
researcher’s contribution to scientific development; finance AI researchers via
contests and competitions; increase Russian-published articles on AI research;
establish transparent methods for research and computer modeling.
(3) Data regulation: Create online repositories to collect, store, and process scientific
data, including for training AI algorithms; provide Russian citizens with tools to
manage their personal data and anonymize personal data posted in the public
domain.
 (4) Hardware and software development: Develop AI-ready microprocessors that run Case of the
faster and use less electricity; simplify the licensing of intellectual rights to AI artificial
development software; create centers for collective AI solutions testing.
intelligence
(5) Education: Use AI to monitor student performance to raise graduation rates; create a strategies
training system for AI professionals in universities and academia; create unified AI open
learning platform for all educational levels; establish a secondary AI vocational
education system; train engineers to allow the creation of domestic AI hardware.
(6) Healthcare: Use AI applications to conduct preventive examinations and to
minimize defects and errors in diagnostic and invasive procedures; integrate AI
tools into the Russian Unified State Health Information System; create National
data bio-bank for big data analysis.

The document presents Russia’s officially recognized definition of AI as “a set of

technological solutions that makes it possible to simulate human cognitive functions [. . .] as
well as to obtain results during the performance of specific tasks that are at least
comparable to the results of human intellectual activity. This set of technological solutions
shall consist of information and communications infrastructure, software [. . .], and data-
handling procedures and services” (The President of the Russian Federation, 2019). The
strategy stresses the strategic importance of AI as a prerequisite for Russia’s entry into the
group of economic world leaders as well as the country’s technological independence and
competitiveness. Even though Russia is not currently considered a leader in the realm of AI,
the document states that Russia has the potential to “[become] an international leader in the
development and use of artificial intelligence technologies”.
This disparity is rather evident in the view of the Unified framework assessment tool:
 Beneficence (1) is a well-represented principle in the Russian AI strategy. Both
Education and Scientific community priority fields call for new funding grants and
programs in AI-training to enhance the overall positive impact of AI on the Russian
system of science and education. The healthcare priority field states that AI could be
vital for early diagnostics and the Russian healthcare system can increase its efficiency
via extensive AI applications, thus significantly raising at least some of the aspects of
people’s life quality. On the basic scale, the beneficence principle in the Russian AI
strategy can be assigned grade “3”. On the explicability scale, this principle is
represented rather poorly – “1”. There are no legal initiatives and all mentioned priority
fields lack project agendas to support their noble goals (Figure 3).
 Non-maleficence (2) is not only extensively mentioned but also declared among the
main principles of the strategy. Legal and ethical priority field calls for the
formulation of the appropriate ethical standards in the field of AI and ethical rules
for human interaction with AI to prevent potentially maleficent use of AI
technologies and ensure their safety for humans. On the basic scale we can assign
grade “3”, but on the explicability scale only grade “1” is possible: Russian strategy
rather eagerly declares support for the non-maleficence principle, but does not
provide any details on how it will be implemented in the form of government-
backed programs and initiatives.
 The autonomy (3) principle is represented poorly. The strategy in review does not
mention any decide-to-delegate models for AI users, simply stating that such
delegation is possible except decisions that might infringe upon the rights and
legitimate interests of individuals. We assign 1 on the basic scale and “0” grade on
explicability scale.
K  Justice (4) is fairly represented in the strategy. The legal and ethical priority field
states that AI use must not infringe human rights in any way as well as be
transparent and non-discriminative. Other priority fields also have some indirect
meaning of ‘just’ AI utilization. Overall, equal access to the benefits of AI is clearly
stated, however, there are not any means and legal initiatives to back this
declaration. We assign” 2” on the basic scale and “0” grade on explicability scale.

The evident lack of explicable support means makes the Russian National AI Strategy looking
overly optimistic. The document assumes that Russian society as well as public organizations and
public enterprises will readily accept the benefits of AI applications and will fully cooperate with
the Strategy’s priorities. Also, Russian AI strategy presumes (and perhaps in a false way) that
relevant Russian federal and regional government institutes will facilitate smooth implementation
of the strategy, fully collaborating with existing public and private AI research projects.

3.3 China
Currently, Chinese companies are the most popular among international tech-investors: in
2018, Chinese Sense Time ($1,200m), UBTech Robotics ($820m), Megvii Technology
($600m), YITU Technology ($300m) and US Dataminer ($391m), CrowdStrike ($200m) and
Pony.ai ($214m) became the leaders in terms of investment in the AI market. The purpose of
AI development in China is to become a “scientific and technical superpower”, a world
center of innovations in the field of AI (the leader in all AI fields) with an active commercial
AI industry and a leader in the creation of industrial robots.
China is one of the most active players in the field of AI, and this applies not only to
technology and market infrastructure development but also to regulation. At present, China has
the most extensive system of legislative acts and state plans containing priorities for the
development of AI-technologies, but this diversity is significantly hindered by the vague and
overly declarative nature of these documents. The need for the development of AI systems is
outlined in the 13th Five-Year Plan for Economic and Social Development of the People’s
Republic of China 2016–2020. According to this document, the Chinese authorities expect
economic growth of at least 6.5% during the five-year period 2016–2020, which is expected to be
achieved through “breakthrough programs” in the field of robotics, innovative technologies for
various purposes, and cybersecurity systems using AI. There is also a need to develop robotic
systems for industrial, service, medical, and other purposes, as well as to promote the use of AI in
the national economy, regardless of the industry (Central Committee of the Communist Party of
China, 2016).

Beneficence
3

1
Non-
Jusce 0
maleficence

Figure 3.
The framework of the
Russian AI strategy Autonomy
principles
Direct representaon Explicable representaon
 However, the most important document declaring China’s desire to become a world Case of the
leader in the development of AI is the Plan for the Development of New Generation AI artificial
Technologies from 2017 (State Council of the People’s Republic of China, 2017). The Plan
outlines a three-step strategy, which the following achievements:
intelligence
(1) By 2020, China will be one of the world’s most advanced levels of AI technologies. By
strategies
2020 AI should have become an important driver of economic growth with the value of
core AI industries to reach 150bn RMB, while that of AI-related industries to reach 1tn
RMB. Chinese enterprises should possess world-leading competitiveness in a series of
technology breakthroughs like in intelligent big data, intelligent cross-media, swarm
intelligence, hybrid enhanced intelligence, and indigenous intelligent systems.
(2) By 2025, China is to become the world’s leading country in the field of some AI
technologies and their applications, and also achieve breakthroughs in
fundamental AI theories. AI should be extensively applied in areas such as smart
manufacturing, smart healthcare, smart cities, smart agriculture, and national
defense, with the value of core AI industries to reach 400bn RMB, while that of AI-
related industries to reach 5tn RMB. The year 2025 will also mark the preliminary
establishment of a new legal framework for AI, including standards, safety
assessments, and supervision.
(3) By 2030, China is to become the world’s leading level in all AI theories,
technologies, and applications. Also, the country will be the global center for AI
technologies and AI economy. The core industry value will exceed 1tn RMB, AI-
related industries value – 10tn RMB. AI will have deep integration in social
governance, national security, and defense.

The Plan indicates six key tasks to implement to fulfill the above objectives, specifically
relating to:
(1) the establishment of an open and cooperative AI technology innovation system;
(2) the cultivation of a high-end and highly-efficient
intelligent economy;
(3) the building of a safe and convenient
intelligent society;
(4) the strengthening of civil-military integration
in the field of AI;
(5) the establishment of a ubiquitous, safe and efficient intelligent infrastructure
system; and
(6) forward-looking layout of a new generation of AI-related major projects.

Despite its detail and consistency, the Chinese strategic plan is perhaps the hardest
document to analyze through a unified framework. Of course, it is declarative no less than
the reviewed US and Russian strategies; however, the concentration of its provisions on
economic and technological advantages brings it in line with business plans, rather than
with government strategic documents:
 Beneficence (1) is a principle that is rarely mentioned but strongly implied in the
Chinese strategic plan. Practically, all strategic plan’s priorities and tasks are meant
to bring economic prosperity to the country for years to come as well as to establish
a new quality of life. That includes the application of innovative AI throughout
education, health care, pension, and other urgent needs involving people’s livelihood.
K There are several key social services in the plan, which should embrace AI technologies
via extensive government support to become more beneficial and effective, namely
Intelligent Education, Intelligent Medical Care, and Intelligent Health and Elder Care
Systems. Still, the plan does not provide any means to achieve this goal. There are no
programs or projects assigned, as well as no legal initiatives to back them. On the basic
scale grade, “3” can be assigned. On the explicability scale – only “1” (Figure 4);
 Non-maleficence (2) is represented in the strategic plan’s guarantee measure 1. Develop laws,
regulations, and ethical norms that promote the development of AI, which calls for safe and
ethical AI use. For this purpose, the document advocates for the development of a legal
foundation for AI as well as for research on AI behavior science. Nonetheless, the strategy
fails to expand on these matters and provides no implementation measures. On the basic
scale we can assign grade “1”, but on the explicability scale only grade “0” is possible.
 The autonomy (3) principle is not represented at all. The strategy in review does not
mention any decide-to-delegate models for AI users. It is stated that AI can enhance
the decision-making process, but any measures to balance AI-assisted decisions are
not mentioned or implied. We assign” 0” on both scales.
 Justice (4) is represented mostly in the meaning of a traceable and accountable AI
deployment (guarantee measure 1. Develop laws, regulations, and ethical norms that
promote the development of AI) and ethical and moral multi-level judgment
structure and human-computer collaboration ethical framework. No further
clarification is provided of these definitions, as well as no supporting means. We
assign” 1” on the basic scale and “0” grade on explicability scale.

4. Conclusions and future directions for research

The Unified framework of five principles for AI in society has proven to be more than just a
coherent theoretical architecture for securing positive social outcomes from AI technology,
but also a potentially viable assessment tool to define how certain strategic documents in the
field define and implement these outcomes. Although social responsibility and ethics in AI
may seem as highly theoretical aspects, they still have a strong practical dimension, which
is unfortunately rarely recognized. At present, there are many voluntary ethical standards of
social responsibility accepted by AI, but they have not yet found broad recognition in
national strategic documents, dealing with AI matters. It often happens that national
strategies in the field of AI seek to address all possible issues at once, but in reality, without

Beneficence
3

1
Non-
Jusce 0
maleficence

Figure 4.
The framework of the
Chinese AI strategy Autonomy
principles
Direct representaon Explicable representaon
a proper ethical framework (of both enabling and constraining nature) they address none at Case of the
all. This is particularly evident if we compare thee documents with one another. A cross- artificial
country comparison diagram of social responsibility principles in AI is presented below
(Figure 5). Composite explicability principle grade here is calculated as the arithmetic
intelligence
average of the country’s basic principles explicability grades. strategies
The analyzed strategies of the US, Russia, and China may lean towards some or other
declared basic principles of the social responsibility, but they fail to explicit them through mutual
cross-priority implementation and accompanying support measures. This evident disproportion
only stresses the significance of the explicability principle, introduced by L. Floridi in the Unified
framework. Without this enabling principle’s proper manifestation, each strategy in the field of
AI becomes just a collection of good wishes. The authors of this paper argue that an effective
national AI strategy should focus not only on declarations but provide a complex framework of
enabling activities and priorities. This holistic approach will allow the nation to unlock the full
potential of AI technologies in an ethical, transparent, and safe way.
As the authors of this paper would like to propose, the convergence of the ethics of AI with
bioethics is not only permissible but desirable. Van Rensselaer Potter’s original concept of
bioethics as an interdisciplinary area of knowledge (Potter, 1994), integrating values of society
with the values of science, can become a blueprint for a similar approach in AI. Therefore, it
seems relevant to consider further similarities between the prospective model of the ethics of AI
and the accumulated wealth of theoretical, methodological, and practical knowledge in bioethics.
Also, it seems relevant to broaden our understanding of the Philosophy of Information in
regards to systems science. If we are going to embrace the crucial concept of Philosophy of
Information, which declares every entity in the Universe as an information object, we must
discover the principles governing the information systems on both physical and metaphysical
levels. And it is the metaphysical level which is of paramount importance since it allows us to
manipulate Shannon information to govern information systems, ourselves included. On this
journey, we must keep in mind that the only coherent metaphysics is that which intends to
build bridges between and unify elements of scientific theories (Ross et al., 2009).

5. Theoretical and practical implications

This research allows us to better understand the linkages, overlaps, and differences between
modern philosophy of information, AI-ethics, social responsibility, and government regulation.
The analysis provided in this paper can serve as a basic blueprint for future attempts to define

USA Russia China

Beneficence
3

Explicability 1 Non-maleficence
Figure 5.
0 Cross-country
comparison diagram
of social
responsibility
Jusce Autonomy principles in AI
K how social responsibility is understood and implied by government decision-makers. The
experience of the USA, Russia and China in this field can be used in a variety of ways to
analyze and improve other countries’ AI development strategies. Certain provisions of the
work may be useful for government agencies and can be applied in developing new
administrative and legal instruments to support the use and development of AI.

References
Albrecht, J.P. (2016), “How the GDPR will change the world”, European Data Protection Law Review,
Vol. 1, pp. 287-289.
Allen, C., Smit, I. and Wallach, W. (2005), “Artificial morality: top-down, bottom-up, and hybrid
approaches”, Ethics and Information Technology, Vol. 7 No. 3, pp. 149-155.
Anderson, R.J. and Sherrock, W.W. (2013), “Ethical algorithms: a brief comment on an extensive”,
available at: www.sharrockandanderson.co.uk/wp-content/ (accessed 03 September 2020).
Arkhipov, V.V. and Naumov, V. (2017a), “Artificial intelligence and autonomous devices in legal. Tr.
SPIIRAN, Vol. 6 No. 47.
Arkhipov, V.V. and Naumov, V. (2017b), “On some issues of the theoretical basis for the development.
Law, Vol. 5 No. 157.
Beauchamp, T.L. and Childress, J.F. (2012), Principles of Biomedical Ethics, Oxford University Press Oxford.
Boddington, P. (2017), Towards a Code of Ethics for Artificial Intelligence, 1st ed., Springer New York.
Borenstein, J. and Pearson, Y. (2010), “Robot caregivers: harbingers of expanded freedom for all?”,
Ethics, Vol. 12 No. 3, pp. 277-288.
Brookings Institution (2018), “A blueprint for the future of AI”, Weapons of the Weak: Russia and AI-
Driven Asymmetric Warfare, available at: www.brookings.edu/research/weapons-of-the-weak-
russia-and-ai-driven-asymmetric-warfare/#footnote-4 (accessed 24 August 2020).
Bynum, T.W. (2010), “The historical roots of information and computer ethics”, in Floridi, L. (Ed.), The
Cambridge Handbook of Information and Computer Ethics, Cambridge University Press,
Cambridge, p. 24.
Central Committee of the Communist Party of China (2016), The 13th Five-Year Plan for Economic and
Social Development of the People's Republic of China (2016–2020).
Cerka, P., Grigiene, J. and Sirbikyte, G. (2017), “Is it possible to grant legal personality to artificial
intelligence”, Computer Law & Security Review, Vol. 3 No. 5, p. 685.
Darnault, C., Parcollet, T. and Morchid, M. (2019) “Artificial intelligence: a tale of social responsibility”,
Association for the Advancement of Artificial Intelligence.
de Mul, J. (2010), “Moral machines: ICTs as mediators of human agency”, Techné: Research in
Philosophy and Technology, Vol. 3 No. 14, pp. 226-236.
Executive Office of the President (2019), Maintaining American Leadership in Artificial Intelligence, E.
O. 13859 of Feb 11, 2019. s.l.: Executive Office of the President, Presidential Document.
European Commission (2018), European Group on Ethics in Science and New Technologies (2018,
March). Statement on Artificial Intelligence, Robotics and ‘Autonomous’ Systems.
Floridi, L. (2002), “What is the philosophy of information?”, Metaphilosophy, Vol. 33 Nos 1/2, pp. 123-145.
Floridi, L. (2008), “Foundations of information ethics [electronic resource]//the handbook of information
and computer ethics”, available at: www.cems.uwe.ac.uk/pchatter/2011/pepi/ (Accessed 01
September 2020).
Floridi, L. (2013), The Ethics of Information, Oxford: Oxford University Press.
Floridi, L. (2019), “What the near future of artificial intelligence could be”, Philosophy and Technology,
Vol. 32 No. 1, p. 1.
Floridi, L. and Cowls, J. (2019), “A unified framework of five principles for AI in society”, Harvard Data Case of the
Science Review, Vol. 1 No. 1.
artificial
Friedler, S.A. Scheidegger, C. and Venkatasubramanian, S. (2016), “On the (im) possibility of fairness”,
arXiv preprint, issue 1609.07236.
intelligence
Future of Life Institute (2015), An Open Letter: research Priorities for Robust and Beneficial Artificial
strategies
Intelligence. s.l, Future of Life Institute.
Future of Life Institute (2017), “Asilomar AI principles”.
Gadzhiev, G.A. (2018), “Whether robot-agent is a person?”, Journal of Russian Law, Vol. 1 No. 29.
Goldman, N., et al. (2013), “Towards practical, high-capacity, low-maintenance information storage in
synthesized DNA”, Nature, Vol. 494 No. 7435, pp. 77-80.
Gurney, J.K. (2016), “Crashing into the unknown: an examination of crash-optimization algorithms”,
Albany Law Review, Vol. 183, p. 79.
Heisenberg, D. (2005), “Negotiating privacy: the European union”, The United States and Personal
Data, Boulder: Lynn Rienner Publishers.
Horvitz, E. and Selman, B. (2009), “Interim report from the panel chairs, association for the
advancement of artificial intelligence”.
House of Lords (2018), “House of lords artificial intelligence committee (2018, April, 16). AI in the UK:
ready, willing and able”.
IEEE (2017), “The IEEE initiative on ethics of autonomous and intelligent systems (2017). ethically
aligned design”.
Kofas, J. (2017), “Artificial intelligence: socioeconomic, political and ethical dimensions. Counter”,
available at: https://countercurrents.org/2017/04/artificial-intelligence-socioeconomic-political-
and-ethical-dimensions/ (accessed 03 September 2020)
McCarthy, J., Minsky, M.L., Rochester, N. and Shannon, C.E. (2006/1955), “A proposal for the dartmouth
summer research project on artificial intelligence, August 31, 1955”, AI Magazine, 15 December,
Vol. 27 No. 4, p. 12.
McCarthy, J., Minsky, M.L., Rochester, N. and Shannon, C.E. (2006 [1955), “A proposal for the
dartmouth summer research project on artificial intelligence, August 31, 1955”, AI Magazine, 15
December, Vol. 27 No. 4, p. 12.
Montreal University (2017), “Montreal declaration for a responsible development of artificial intelligence, s.
l.: announced at the conclusion of the forum on the socially responsible development of AI”.
Morkhat, P.M. (2017), Artificial Intelligence: A Legal Perspective, Moscow: Buki Vedi.
Morkhat, P.M. (2018), “Concerning the question about the legal personality of electronic person”, Legal
Studies, Vol. 4 No. 1.
Noorman, M. and Johnson, D.G. (2014), “Negotiating autonomy and responsibility in military robots”,
Ethics and Information Technology, Vol. 16 No. 1.
O’Heigeartaigh, S. (2013), “Would you hand over a moral decision to a machine? Why not? Moral
outsourcing and artificial”, available at: http://blog.practicalethics.ox.ac.uk/2013/08/would-you-
hand-over-a-moral-decision-to-a-machine-why-not-moral-outsourcing-and-artificial-intelligence/
(accessed 03 September 2020).
Partnership on AI (2018), “Partnership on AI (2018). Tenets”.
Pavaloiu, A. and Kose, U. (2017), “Ethical artificial intelligence–An open question”, arXiv preprint,
issue arXiv:1706.03021.
Pavaloiu, A. and Köse, U. (2017), “Ethical artificial Intelligence – an open question”, Journal of
Multidisciplinary Developments, No. 2, pp. 15-27.
Potter, V.R. (1994), “An essay review of – global responsibility. In”, Perspectives in Biology and
Medicine, Vol. 37 No. 4, pp. 546-550.
K Rabinow, P. (1996), Essays on the Anthropology of Reason. c1996 Xvii ed, Princeton, N.J.: Princeton
University Press.
Ross, D., Spurrett, R. and Collier, J. (2009), Every Thing Must Go: Metaphysics Naturalized. 1st ed. s.l,
Oxford University Press.
Schlenoff, C.I. (2015), “IEEE standard ontologies for robotics and automation. IEEE std 1872-2015”, 10
April, pp. 1-60.
Shannon, C.E. (1948), “A mathematical theory of communication”, Bell System Technical Journal,
Vol. 27 No. 3, pp. 379-423, 623.-656.
Sharkey, N. (2008), “The ethical frontiers of robotics”, Science, Vol. 322 No. 5909, pp. 1800-1801.
SIPRI (2019), “The impact of artificial intelligence on strategic stability and nuclear risk”, available at:
www.sipri.org/sites/default/files/2019-05/sipri1905-ai-strategic-stability-nuclear-risk.pdf (accessed
24 August 2020).
Stanford Encyclopedia of Philosophy (2019), “Stanford encyclopedia of philosophy, definition of agency”,
available at: https://plato.stanford.edu/entries/agency/#::text=In%20very%20general%20terms%
2C%20an,a%20standard%20theory%20of%20action. (accessed 24 August 2020).
State Council of the People’s Republic of China (2017), “Next-generation artificial intelligence
development plan”.
Sullins, J.P. (2006), “When is a robot a moral agent?”, International Review of Information Ethics, Vol. 6 No. 12,
pp. 23-30.
The President of the Russian Federation (2019), Decree of the President of the Russian Federation on the
Development of Artificial Intelligence, s.l, Official Portal of Russian Legal Information.
UNESCO (2018), “Towards a global code of ethics for artificial intelligence research”, The UNESCO
Courier, Vol. 3 No. July-September 2018.
United States Government. (2019), “National AI R&D strategic plan: 2019”.
Wiener, N. (1961), Cybernetics or Control and Communication in the Animal and the Machine, Reissue of
the 1961 Second Edition. 2nd ed. s.l, The MIT Press.

Further readings
Grishin Robotics (2017), “Federal act’ on amendments to the civil code of the Russian Federation in terms
of improving the legal relationship management in the field of robotics” (in russian)”, available at:
http://robopravo.ru/uploads/s/z/6/g/z6gj0wkwhv1o/file/bESvQz3Y.pdf (accessed 28 August 2020).
Kant, I. (1999), Critique of pure reason (the cambridge edition of the works of immanuel kant)”,
Translated and edited by Paul Guyer and Allen W. Wood. s.l, Cambridge University Press.
Robopapravo (2017), “The initiative for study legal aspects of robotics and artificial intelligence
‘robopapravo’. [online]”, available at: https://robopravo.ru/matierialy_dlia_skachivaniia#ul-id-4-
35 (Accessed 28 August 2020).
Vasilyev, A.A., Ibragimov, Z.I. and Gubernatorova, E.K. (2019), “The Russian draft bill of ‘the Grishin
law’ in terms of improving the legal regulation of relations in the field of robotics: a critical
analysis”, Journal of Physics: Conference Series, Vol. 1333, p. 052027.

Corresponding author
Anton Saveliev can be contacted at: anton.saveliev@gmail.com

For instructions on how to order reprints of this article, please visit our website:
www.emeraldgrouppublishing.com/licensing/reprints.htm
Or contact us for further details: permissions@emeraldinsight.com

View publication stats