A Methodology for Ethics-by-Design AI Systems: Dealing with Human Value Conflicts

A Methodology for Ethics-by-Design AI Systems: Dealing with Human Value Conflicts

Abstract

The introduction of artificial intelligence into activities traditionally carried out by human beings produces brutal changes. This is not without consequences for human values. This paper is about designing and implementing models of ethical behaviors in AI-based systems, and more specifically it presents a methodology for designing systems that take ethical aspects into account at an early stage while finding an innovative solution to prevent human values from being affected. Two case studies where AI-based innovations complement economic and social proposals with this methodology are presented: one in the field of culture and operated by a private company, the other in the field of scientific research and supported by a state organization.

artificial intelligence, ethics, value, ethics-by-design, AI applications

I Introduction

Artificial intelligence (or “AI”) is often defined as the simulation on a machine of so-called “intelligent” processes. Being both an applied and theoretical field, this discipline of computer science is covering the range from weak AI (can machines act intelligently?) to strong AI (can machines really think?) [29]. For the past decade, the first approach has brought AI back to the forefront, especially with the development of new machine learning techniques, such as deep learning models [35]. This technique has created extremely effective AI applications in the area of pattern recognition or information selection problems for decision making, with programs that extract information from raw data and learn to improve their skills from existing examples. With this learning process, AI systems can perform complex tasks in place of humans.

However, the arrival of this new technique has brought a number of ethical issues. Firstly, artificial intelligence programs reason in a simplistic way, but the real world is complex and full of unexpected events, which the machine has great difficulty in dealing with. Secondly, when an AI program learns about data collected on past situations, it performs statistical deductions and transforms correlations between variables into implication relationships. This can lead to problems with dramatic consequences such as gender bias in resume analysis support system by denying women access to managerial positions [8], or racial bias in legal decision support system to predict future criminals [3].

The concern for the developers is to optimize some specific criteria, e.g., efficiency and usability. For instance, online e-commerce marketplaces seek to optimize the relevance of the connection between supply and demand. They collect a maximum of information from consumers and suppliers, and use a whole set of AI-derived strategies to maximize the matching between the different parties. Attention-economy-based companies seek to keep their users on their websites and apps for as long as possible to offer them advertising. The collection of personal data leads to personalization which may have certain advantages, such as recommending more relevant content, but also important problems relating to the preservation of privacy, attacks on democracy [37], or the limitation of access to diversity [26]. Intelligent system algorithms are black boxes that are impossible to understand, they are unregulated and difficult to question in the case of the presence of bias, and in some cases they amplify inequalities [25].

The highlighting of these problems has led to reactions from States and civil society. Some initiatives, for example in the USA (Asilomar AI principles,1 “AI Now” Institute…), in Canada (Montréal Declaration for a responsible development of Artificial Intelligence2), or in France (Villani Report “For a meaningful Artificial Intelligence”3), with the production of reports, charters, lists of principles or laws, are moving in the direction of taking ethical concerns into account early in the development of AI systems. However, there is not much information provided on how to make this possible.

{comment}

For designing and implementing models of ethical behaviors, this paper proposes in the following a methodology for building ethics-by-design AI systems. This methodology is explained and illustrated by means of two examples of innovative applications carried out during PhD studies conducted under the supervision of the author and how these AI applications are able to respond to problems of human value conflicts.

For designing and implementing models of ethical behaviors, this paper proposes in the following a methodology for building ethics-by-design AI systems. This methodology is explained and illustrated by means of two innovative applications developed during PhD works supervised by the author. We will explain how these AI applications are able to respond to the problem of human value conflicts.

{comment}

In the following, for designing and implementing models of ethical behaviors, I propose a methodology for building ethics-by-design AI systems. To show how this methodology works, I will present two examples of innovative intelligent system applications carried out during PhD thesis studies carried out under my supervision and how they are able to respond to problems of human value conflicts.

Human values: what is important to people in their lives, with a focus on ethics and morality.

As descriptive, ethics attempts (a) to present a typology of what all human communities experience, in varying degrees and with varying content, as the objects of moral conscience, that is, the values themselves and their order of relative worth. It must also (b) describe the elements of moral consciousness, specifically the noetic structure whereby human beings become aware of values and make them function in their moral judgments.

As normative, ethics aspires (a) to a critique the various ethea of cultures, specifically their standards of right and wrong, and to the establishment of a theory of values that permits a correct evaluation of its norms. Further, it seeks (b) to characterize the nature of virtue and the contents of the virtues, and to establish a rank of relative value among them. Thus in its two functions, ethics must make explicit the elements of moral consciousness and its objects, and must establish a coherent order among those norms as criteria of right behavior (as in systems of law or moral teachings).

Ii Methodology for Ethics-by-Design AI

The main idea of the three-step methodology proposed here is to take advantage of collective intelligence. The introduction of technologies derived from artificial intelligence leads to risks of undermining values for which the different stakeholders may have contrasting views.

Ii-a First Step: Innovations in an Economic and Social World

The first step of the methodology is to set up an economic and social model respectful of human values where the contribution of AI technologies can be integrated as a complement to the skills of human beings. The quote “They who can give up essential Liberty to obtain a little temporary Safety, deserve neither Liberty nor Safety” attributed to Benjamin Franklin makes us think that certain values are necessarily contradictory: liberty/privacy vs. security/safety. {comment}Even though information and communication technologies (ICT) are often seen as privacy-invading technologies, privacy by design approaches and technologies try to guarantee this fundamental human right [14]. When resources are limited, choices need to be made. These choices will necessarily follow certain values and favor certain behaviors supported by some groups of individuals over others.

However, in the case of adding a new technology such as AI, we must find a way not to optimize a single value at the expense of others. It is important to try always to find an alternative and innovative solution that guarantees respect for all human values. For example, in our study made in the field of predictive judicial analytics [24], we came to the conclusion that the values of knowledge and trust are essential to give credit to decisions produced by intelligent systems. If the deep learning models cannot guarantee understanding by specialists (i.e., lawyers), they must be replaced by other more comprehensible models (rule-based approaches, Bayesian methods), otherwise these AI models will no longer be able to contribute to their role as a tool to serve justice and help defend litigants.

Ii-B Second Step: Ethical Matrix

The methodology continues with the clarification of the main values involved in the introduction of a new AI system for all stakeholders by setting up a decision-making tool called the “ethical matrix.” This tool was initially proposed by Ben Mepham to facilitate judgements on bioethic questions [21], in order to identify the values that are threatened. {comment} (e.g., the use of genetically modified fishes, policy interventions in the obesity crisis, or animal sentience concerns) By focusing more specifically on a small number of ethical principles (e.g., respect for wellbeing, for autonomy and for justice) on a given subject (e.g., the impact of new technologies in food and agriculture), it is possible to elicit the problems and concerns of different stakeholders or interest groups. The selected ethical principles constitute values that form the columns of the matrix. The rows of the matrix consist of each stakeholder caught up with the issue in question. Each cell of the matrix specifies the main criterion to be satisfied for a stakeholder for a given principle. {comment}So, the structure of the ethical matrix consists of stakeholders on the y-axis, principles on the x-axis, and questions and answers in their intersections. The ethical matrix can then be seen as a checklist of concerns, structured around established ethical theory, but it can also be used as a means of provoking structured discussion between the different interest groups.

One of the most important difficulties of the ethical matrix approach concerns the choice of the values to use. A value is a way of being and acting for a person or a group of people, and people consider that certain behaviors resulting from a value are more desirable than others. From Antiquity (e.g., with Plato) to the present day, authors have proposed lists of fundamental values, such as the Axiology –defined as the philosophical science of values– proposed by the philosopher Max Scheler [31], or the “Theory of Basic Human Values” proposed by the social psychologist Shalom H. Schwartz [32]. {comment} Questions about values have been asked since Antiquity. For example, for Plato, if the basic values are the truth, the beauty, and the goodness, the cardinal virtues are the prudence (also translated by the restraint or the moderation), the justice (or the fairness), the temperance (or the wisdom) and the courage (or the fortitude). Since then, many authors have proposed models, theories and lists of values, such as the Axiology –defined as the philosophical science of values– proposed by the philosopher Max Scheler [31], or the “Theory of Basic Human Values” proposed by the social psychologist Shalom H. Schwartz [32]. In the French book Petit Traité des Valeurs [9] (“Little Treatise of Values”), 35 values are proposed (Table I) and studied in a philosophical point of view (e.g., could things be beautiful, useful, interesting or pleasant in the absence of anyone to appreciate or experience them?)

art authenticity beauty childhood competency
confidence creativity dignity equality eroticism
freedom friendship health honor humbleness
humor impartiality interesting justice knowledge
life love luxury pleasure power
property privacy righteousness sacred solidarity
sublimate tasty tradition usefulness wellbeing
TABLE I: Ethical Values

The arrival of artificial intelligence highlights a specific value: the efficiency. By taking into account a greater number of variables on a greater number of examples that no human being is capable of accumulating in a lifetime, intelligent programs are able to accomplish very complex tasks brilliantly and effortlessly. {comment}Reenacting a new combat between man and machine, AI is considered as a new avatar of these world-changing inventions such as the loom, the steam engine or other productivity-improving technologies, generating as many fears as hopes like any revolution. However, this efficiency, introduced in a disruptive manner in fields which until now felt protected, such as medicine or justice, seems to be predatory for other values.

Faced with this difficult task, it seems important to us to select a list of basic values, for example the 10-item list of values selected by the Montréal Declaration for a Responsible Development of Artificial Intelligence [23]: well-being, respect for autonomy, protection of privacy and intimacy, solidarity, democratic participation, equity, diversity inclusion, prudence, responsibility, and sustainable development. However this list is neither necessary nor exhaustive. Depending on the application context, values more likely than others to be affected by an AI technology must replace certain values among the 10 or be added to this list.

{comment}

There is thus a diversity of scale of values in the world population, on which it is difficult to give an opinion or a judgment, but which it is necessary to take into account. This is why a principle, such as the 11 in the list of 23 Asilomar Principles (“Human values: AI systems should be designed and operated so as to be compatible with ideals of human dignity, rights, freedoms, and cultural diversity”), seems difficult to implement.

Depending on cultures and countries, some values are put forward more than others, for example “liberty, equality and fraternity” is the official motto of the French Republic, while other countries insist on a value such as faith (e.g., the United States of America: “in God we trust”). Furthermore, the values are not the same at the personal level (with a set of instrumental or terminal values in the Rokeach Value Survey like capability, imagination, freedom or pleasure), at the level of a small group of individuals or even the level of a state (with values such as democracy, equity, legality and accountability) [28].

It is not always possible to formulate a hierarchy of values explicitly, but the relative importance of values appears when faced with situations in the real world or in role plays. In Moral Machine online test, authors have conducted a survey from January 2015 to December 2019 collecting the preference data from 1.3 million people by choosing between two evils (e.g., running over pedestrians or sacrificing themselves and their passenger to save the pedestrians?) in the event of an accident involving an autonomous vehicle [6]. It emerged from this study of people from 233 different countries and territories that individuals, whatever their nationality or culture, shared three main characteristics: (1) humans are favored over animals, (2) the minimum number of victims is preferred, and (3) the young people are more saved than the elderly people. But there are cultural variations according to geographic areas. For example, in France, women are more willingly saved than in the United States. In Asia and the Middle East, the elderly people are more spared and social status is less taken into account, while in Latin America, Central Europe or France, young people are spared more and people with low social status are disadvantaged.

According to Friedman et al. (2013), human values often implicated in system design [13] are human welfare, ownership and property, privacy, freedom from bias, universal usability, trust, autonomy, informed consent, accountability, courtesy, identity, calmness, and the environmental sustainability.

Fort de ces habiletés techniques dans tout un ensemble de domaines où l’esprit humain se croyait supérieur et protégé, l’intelligence artificielle vient modifier les règles du jeu, aussi bien dans le domaine de la connaissance, de l’accès aux sources d’information.

With these technical skills in a whole range of areas where the human spirit believed itself superior and protected, artificial intelligence comes to modify the rules of the game, as well in the field of knowledge, of access to sources of information.

{comment}

As Stuart Russell mentions [30], from the earliest days of cybernetics, Norbert Wiener –the cybernetics’originator– expressed his concerns about what we introduce into a machine: “If we use, to achieve our purposes, a mechanical agency with whose operation we cannot efficiently interfere (…), we had better be quite sure that the purpose put into the machine is the purpose which we really desire.”

But how is ‘we’?

The values of individuals can differ greatly from one individual to another, or from one social group to another. It is then recommended to take into account all possible stakeholders, even those who cannot express themselves (e.g., animals, nature). In the field of artificial intelligence, the stakeholders refers especially to researchers, developers, manufacturers, providers, policymakers, and users [10]. In the generic ethical matrix manual [20], a protocol has been proposed for these stakeholders to meet in workshops in order to take into account the diversity of opinions.

Ii-C Third Step: Value Sensitive Design Approach

In the last step, the plural visions and their consequences on some human values provided by the matrix allow to draw the guidelines for integrating human values throughout the design process and this development by following the principles of the “Value Sensitive Design” (VSD) introduced by Batya Friedman and her colleagues in the late 1980s [12]. The VSD theory has influenced the Dutch approach to responsible innovation [36]. In this approach, human values are integrated throughout the design process and this development is done using a three-phase survey: conceptual, empirical and technological [13].

There are many different methods of applying this theoretical approach in practice to real-world problems [11]. Among the advice given by the authors, they recommend starting with a value, a technology or a context of use. Then they propose to identify the direct and indirect stakeholders, as well as the benefits and harms for each stakeholder group. The benefits and harms are then mapped onto the corresponding values, allowing the identification of potential value conflicts.

{comment}

The VSD theory has influenced the Dutch approach to responsible innovation and value sensitive design. In [36], some examples are given of design for the values of accountability and transparency, design for the values of democracy and justice, human well-being, inclusiveness, presence, privacy, regulation, responsibility, safety, sustainability, and trust in various domains (alphabetically, from “agricultural biotechnology” to “water management”).

Iii Case Study 1: Music Recommender Systems

With the advent of the Internet, it was predicted that this would be an opportunity for musical creation aimed at a niche audience. Unfortunately, it did not happen. {comment}The public was going to be able to have access to a greater diversity of production by musical artists, and these, even if they offered music of interest only to a niche audience, were going to be able to more easily find their audience. Apart from a few notable exceptions, access to music via the Internet has only strengthened trends: popular music, artists or genres have become even more popular, and lesser-known artists have found it even more difficult to find their audience, producing an impoverishment of musical diversity [7].{comment}: worldwide, according to the age group, everyone listens to roughly the same thing, with a little variety depending on their birth tongue and native culture. Yet we have varied musical desires and changing and evolving musical tastes, with a certain appetite for musical discoveries.

One of the reasons for the lack of diversity is linked to what Eli Pariser calls the “filter bubble” [26]. Personalized recommendations are based on the processing of big data: the algorithms recommend to individuals products that are similar, or products appreciated by individuals sharing common traits, or popular products. With machine learning, recommendation algorithms will identify people’s profiles, extract the most prototypical traits, and make only recommendations corresponding to a caricature of everyone’s interests. The modus operandi of the recommender systems may be suitable for mainstream consumers who want to go directly to the greatest hits. However, this “superstar effect” –where the winner takes all [7]– impoverishes musical diversity and does not allow musical niches to emerge, ending up stifling a musical creativity that can no longer reconnect with its audience.

If the digital music market is held only by major record labels supporting a musical offer limited to a few superstars, are the music and musical artists produced by independent record labels destined to disappear? In addition, we have become accustomed to paying the container (e.g., a subscription to an Internet service provider, a nice smartphone) but not the cultural content as such (e.g., music and videos on online video-sharing platform, mobile applications on app stores). As the new digital use is mainly free, how can we make people accept to pay to access to unknown musical content produced by indie labels?

According to the methodology proposed here, the first step consists of a response from the economic and social world. Such a response exists in the form of a company offering the general public innovative solutions to discover independent creations. The French social economy company 1Dlab4 has decided to take up this challenge. To be able to give an answer to the question of how to support the varied musical creation produced by independent record labels, this company has found a two-level solution.

The first level solution is economic. The company proposes a fairer remuneration model to the creators (artists and content producers). The business plan works according to a business-to-many model (B2M). The company provides a music streaming platform (1DTouch, diMusic) but also works in the physical world to build a network of places and actors for favoring new ways of sharing, meeting, discovering and creating a cultural diversity. The streaming platform access is possible by a subscription fee paid by the project located partners (libraries, showrooms, work’s councils, administrative divisions…) to their members. {comment}The second level solution is provided by the work of Pierre-René Lhérisson during his doctoral studies conducted under the supervision of Pierre Maret and myself: it is a technical innovation allowing the general public to discover pleasant independent creations with bold recommendation tools that we have designed following an ethics-by-design approachThe second level solution was proposed in the PhD work of Pierre-René Lhérisson, supervised by Pierre Maret and the author of this paper: it is a technical innovation allowing the general public to discover pleasant independent creations with bold recommendation tools that has been designed following an ethics-by-design approach [17].

The views on values such as music (or more generally to culture as a whole), equity, diversity, trust and usefulness are very different depending on whether the stakeholders are the consumers (who listen to music), the creators of cultural contents (artists), the media services providers or the independent record labels. Following the second step of our methodology, we build an ethical matrix (Table II) allowing to see how to develop an innovative solution capable of meeting all the criteria.

Respect for: Culture (music) Equity Diversity Trust Usefulness
Consumers (Listeners)
Have access
to a maximum
of music
Have the same chance
to access interesting music
and make new
musical discoveries
Have access to a
maximum of diverse
music genres
and music artists
Receive music recom-
mendations adapted to
musical tastes and desires
Access interesting
and varied music at
a lower cost with
personalized recommendations
Creators of Cultural Goods (Artists)
Have access to a maximum
of potentially interesting
cultural sources
Have the same chance
of being recommended,
regardless of popularity
Be recommended to
various listeners likely to
appreciate their creations
Believe in the fairness
of the music recommendations
Be paid in order to
make a living from
their creations
Media Services Provider
Have the richest
and most varied
music catalog
Propose recom-
mendations so that
all artists can find
their audiences
Propose a varied offer
intended for specialized
and amateur listeners
Have professional quality
music to offer customers
or subscribers
Offer a user-friendly music
platform with an inexpensive
solution for customers
but allowing artists to
make a living from their work
Indie Labels
Discover new talents;
Support artists’ profes-
sional development paths;
Facilitate meeting
with the audience
Give all
artists the
same chances
of success
Not concerned
(an indie label is
often specialized in a
particular musical genre)
Promote quality music;
See all of their
artists appear in
the recommendations
Introduce a potentially
interested public to the
label or a new music genre
TABLE II: Ethical Matrix of AI-based Model of Fair Music Recommender System in an Indie Streaming Platform.

In the third step of the methodology, we follow an approach inspired by the VSD theory to integrate early in the design the values defined in the previous step for the development of a fair tool for recommending indie content.

First of all, we developed a distance measure between musical items (music, songs, artists, music genres) based on different criteria (text information of the music, properties extracted from the sound signal, musical classification performed by experts) as well as on the perceptual distance sensation obtained through an experiment with listeners{comment} [18]. Then we proposed an “optimal diversity function” using this multi-criteria distance measure in order to find a compromise between the values diversity and efficiency (or usefulness). {comment}To establish this function, we were inspired by the Mexican hat wavelet, a function which is calculated quite simply because it is the negative normalized second derivative of a Gaussian function. As an example of use, let us mention the computation of the connection strength in a self-organizing map (SOM), an artificial neural network model introduced by Teuvo Kohonen in the 1980s [15]. For the SOM, the Mexican hat is used to amplify the contrasts for neurons distributed in networks of a single layer on a flat map and the lateral connections are used to create a competition between neurons. The lateral connections are used to produce excitatory or inhibitory effects, depending on the distance to a given neuron. The Mexican hat function describes synaptic weights between neurons in the Kohonen layer: the higher value of the connection is close to the neuron (the top of the sombrero), acting as an excitatory effect, then the values decrease as it moves away from the neuron until they become negative (playing an inhibitory effect). To get this optimal diversity function, we projected the distances between items on the (negative) Mexican hat wavelet function to establish clear distinction between similar, dissimilar, and sufficiently dissimilar items (Figure 1).5

Fig. 1: Mexican hat and optimal diversity function. The optimal diversity is obtained when the items are not perceived too similar nor too dissimilar.
{comment}

is negative for values too close to the item (poor diversity), then it becomes positive when we move away from the item to a maximum value, and decreases until it becomes zero if the distance to the item is too important.

For example, a listener enjoys trip hop music, a musical genre resulting from the fusion of hip hop and electronic music. Based on this information, it is unreasonable to recommend music located at other ends of the spectrum of musical genres (e.g., classical music or hard rock). On the contrary, it is more relevant for the listener to recommend slightly different items, such as music of the acid house, indie soul or dub type, depending on whether the musical orientation is respectively rather in the direction of jazz, funk music or reggae (Figure 2).

Fig. 2: Musical genres and optimal diversity. Compared to a given item (here, trip hop genre), the diversity will be optimal when it is far enough from the item in question, but not too much (e.g., acid house, electro jazz, dub or indie soul can be recommended genres because the diversity is optimal).

Each item (music, song, artist or musical genre) is located in a multidimensional space where it is always possible to find other items at more or less close distances, therefore of a more or less optimal diversity value. In order to guarantee respect for the equity value, it is thus possible to orient the sense of diversity towards basins of attraction where artists are less recommended than others in order to make them discovered.

To maintain the trust of the user to the recommendations made by the streaming platform, the system indicates to the user that the proposed recommendations are bold, i.e., they are not targeting to the music that the listener is used to listening to, but that there is a chance that the user may still like it.

If, after having tried these recommended musics in a daring way, the listeners express their dissatisfaction, the personalization can adapt by decreasing the distance of optimum diversity and the system can thus propose musics closer to what they already appreciate. If, on the other hand, the listeners have eclectic tastes and want bolder discoveries, the distance of optimum diversity can increase for these people.

Finally, by taking into account the opinions of the various players in a music platform system at a very early stage, it is possible to propose an innovative solution guaranteeing respect for a priori contradictory ethical values.

Iv Case Study 2: Research-paper Recommender Systems in Scientific Digital Libraries

Access to knowledge is a major problem for human beings. This issue affects populations of all ages and at all levels: education for all the children of the planet, fight against (digital) illiteracy, and access to expert knowledge for all scientists, etc. In the latter case, there are notable initiatives that are part of an open science perspective, but articles published in scientific journals or renowned conference proceedings still remain mainly the property of scientific publishers and their access is most often made by purchasing articles or by subscribing to the publisher’s platform. Yet this access to knowledge is necessary for the production of knowledge because researchers are science prosumers: {comment}when researchers are not involved in other tasks required by their position, researchers “consume” and “produce” knowledge, and this activity takes one of the two following paths: exploration or exploitation{comment} [5]. In the exploitation phase, researchers use their existing knowledge to produce new knowledge by following a discovery process specific to their discipline. In the exploratory phase, researchers update their knowledge by discussing with colleagues in their laboratory, reading books or scientific articles, or attending seminars and conferences. Exploratory research is an essential phase of researcher’s activity [19], and exploitation and exploration are two complementary phases that feed on each other.

In addition to the cost of accessing these sources of knowledge, even if these scientific digital platforms are multidisciplinary, the way they are designed and the means they offer to access interesting information will not promote the knowledge transfer between disciplines. Each scientific discipline has its own vocabulary, its own jargon. Without knowing the good keywords to enter in the search engine of the digital library platform, scientists thus find themselves isolated in a “filter bubble” [26] preventing them from accessing the content of articles from other disciplines. However, results can be obtained when the research transcends the limits of a discipline. It is well established that exchanges (knowledge transfers) between disciplines are very fruitful [16].

Faced with these problems, {comment}the Scientific and Technical Information Department (DIST) of the French state research organization (CNRS) has proposed a solution that is part of the social and economic world –as the first step in our methodology– by launching an ambitious project in two parts. The first part concerns the creation of a scientific digital library accessible to all French research institutions called ISTEX6 [33], resulting from a massive resource acquisition policy, obtained by entering into operating contracts with a lot of international and French-speaking publishers. {comment}The second part concerns the exploitation of these resources in scientific literature by funding projects making use of techniques applied to the text, and in particular text and data mining [34]. During the PhD studies of Hussein Al-Natsheh that I supervised with Djamel A. Zighed, I had the opportunity to lead a 4-people team working on such a project with the following goal: to provide assistance to users of ISTEX resource by offering recommender systems that promote disciplinary diversity. The second part concerns the exploitation of these ISTEX resources with text and data mining techniques [34]. Together with the PhD student Hussein Al-Natsheh and with Djamel A. Zighed (co-supervisor), we had the opportunity to lead a 4-people team working on this goal: provide assistance to users of ISTEX resource by offering recommender systems that promote disciplinary diversity.

For researchers (whether they produce or seek knowledge), scientific publishers, research organizations and public authorities, interests diverge greatly, and their respective views on values such as knowledge, equity, diversity, trust or usefulness do not involve the same issues, which can lead to conflicts, as shown on the ethical matrix (second step: Table III).

Respect for: Knowledge Equity Diversity Trust Usefulness
Knowledge- producing Researchers
Want their work to
be accessible to the
scientific community
Have the same
chance of reaching an
interested readership
Can be read even by
people not from the
same scientific discipline
Be able to be found
when a reader is
interested in their work
See their work
appropriately
recommended
Knowledge- seeking Researchers
Have access to a
maximum of potentially
interesting knowledge resources
Not to be limited
in their search
for information
Have access to
various sources
of knowledge
Believe in the
seriousness of the
recommended publications
Access useful
information at low cost
Scientific Publishers
Have the most attractive
knowledge offer in
the digital library
Not concerned
Propose a varied
offer intended for
specialized communities
Have quality publications
to offer to customers
or subscribers
Find a cost
effective publishing
solution
Research Organizations and Public Authorities
Participate in the increase of
knowledge at national level
Give all research
stakeholders access to
information sources
Give all scientific
communities the same
chances of being valued
Promote trustful and
quality research
Have an effective and
inexpensive solution
TABLE III: Ethical Matrix of AI-based Model of Fair Research-Paper Recommender System in a Scientific Digital Library.

In the third step of the methodology, we integrate the values of the ethical matrix into the design of an innovative solution, following the VSD approach. Here, we focus on the value of diversity which is threatened by the value of a precision-based usefulness. As part of the projects funded by the CNRS and intended to benefit from the ISTEX platform, we have proposed to develop a research-paper recommender system whose purpose is to favor diversity. Instead of being focused on the sole criterion of accuracy of results, as most other systems do [4], we propose a system able to recommend scientific papers through a semantic similarity model [1] based on computational linguistics methods and word embedding techniques [22], as shown on Figure 4.

Fig. 3: Semantic-similarity based recommendation of scientific papers. First, a vector representation of the scientific terms is created with word embedding techniques from the contents of the research articles of the library (represented by white dots in a multidimensional space). Second, the interesting papers provided by the user are projected in the vector representation space (represented by red dots) to define a target area of interest. Third, the neighboring articles of the target (black dots) in the vector representation space are activated to recommend diverse items, sometimes coming from other disciplines, but connected through semantic links to the input papers.

In this system, the user does not enter a set of keywords but a list of articles dealing with a given subject of interest. The content (abstract, plain text) of these articles is projected in a multidimensional space constructed by the text representation of the scientific papers of the multidisciplinary digital library (blue arrow in Figure 4). As the representation space is constructed by semantic similarities resulting from a proximity of the presence of terms in the texts, the more or less distant neighborhood will bring articles to recommend (arrows in green) comprising more or less diversity with the input articles. In the experiments we launched, we collaborated with sports science researchers who were working on a test to detect motor skills. Thanks to our system, we have been able to recommend research papers to the sports science researchers –deemed relevant by them– from different disciplines such as psychology, linguistics and defectology (e.g., studies on sign language used by deaf people, another form of complex movements with meaning). This kind of transfer of knowledge between disciplines has hitherto never been established.

By examining in detail the different values involved using the ethical matrix, it was possible to explain them better and to realize that the usefulness of the system does not necessarily imply the quest for precision: diversity also matters. The early integration of the diversity value in the design of the research-paper recommender system allows researchers in the exploratory phase to have access to knowledge from various sources. {comment}Through the pooling of resources, access to knowledge is equitable for all. ISTEX platform allows researchers to be treated fairly: regardless of the financial resources of the institutions to which they belong, they can access to the same sources of knowledge. With this recommender system, ISTEX platform promotes scientific diversity and allows all scientific disciplines to be represented with equity and all their articles to be recommended. From an epistemological point of view, this diversity is essential because otherwise a researcher using a scientific digital library will only have access to sources of information from his/her own discipline. The result will be that the documents found will only confirm the initial point of view of the researcher using this digital library, thus calling into question one of the golden rules of science which is the property of “falsifiability” [27].

{comment}

###############################################################################################

The Scientific and Technical Information Department (DIST) of the French National Centre for Scientific Research (CNRS) has decided to tackle this problem by launching an ambitious project in two parts. The first part concerns the creation of a scientific digital library accessible to all French research institutions: the project ISTEX7 [33], resulting from a massive resource acquisition policy, obtained by entering into operating contracts with a lot of international and French-speaking publishers. The second part concerns the exploitation of these resources in scientific literature by funding projects making use of techniques applied to the text, and in particular text and data mining (TDM) [34].

As part of one of these TDM projects, we proposed a research-paper recommender system whose purpose is to favor the diversity (i.e., the recommendations of scientific articles from other disciplines) by the discovery of new keywords and by the extension of the results through a semantic similarity model [1, 2]. The problem of conventional research-paper recommender systems [4] is that they focus almost exclusively on the criterion of relevance of the results. This improves the precision performance measure, but even if it does not have too much effect on the recall measure, it greatly affects the diversity of results. Other techniques have tried to add some noise to the list of recommended results (e.g., by random walk methods), but it is rare that the papers obtained are interesting by operating only with chance. In order to obtain diversity in the results and to favor the chance of obtaining serendipitous discoveries, we have proposed a semantic-related similarity model to recommend scientific papers with computational linguistics methods and word embedding techniques [22], as shown on Figure 4.

Fig. 4: Semantic-similarity based recommendation of scientific papers. First, a vector representation of the scientific terms is created with word embedding techniques from the contents of the research articles of the library (represented by white dots in a multidimensional space). Second, the interesting papers provided by the user are projected in the vector representation space (represented by red dots) to define a target area of interest. Third, the neighboring articles of the target (black dots) in the vector representation space are activated to recommend diverse items, sometimes coming from other disciplines, but connected through semantic links to the input papers.

In the research-paper recommendation model we proposed, the user does not enter a set of keywords into the system, but a list of articles dealing with a given subject of interest. The content (abstract, plain text) of these articles is projected in a multidimensional space constructed by the text representation of the scientific papers of the multidisciplinary digital library (blue arrow in Figure 4). As the representation space is constructed by semantic similarities resulting from a proximity of the presence of terms in the texts, the more or less distant neighborhood will bring articles to recommend (arrows in green) comprising more or less diversity with the input articles.

ISTEX platform allows researchers to be treated fairly: regardless of the financial resources of the institutions to which they belong, they can access to the same sources of knowledge. Thanks to this recommender system, the platform promotes the scientific diversity and allows all scientific disciplines to be represented with equity and all their articles to be recommended. From an epistemological point of view, this diversity is essential because otherwise a researcher using a scientific digital library will only have access to sources of information from his/her own discipline. The result will be that the documents found will only confirm the initial point of view of the researcher using this digital library, thus calling into question one of the golden rules of science which is the property of “falsifiability” [27].

{comment}

— cf. note proche de l’appel de la table 2. Il faudrait ici faire le bilan de l’usage de table 2 sur le résultat final obtenu. –¿ la réfléxion sur les critères éthiques autres que la précision permet de faire concevoir et d’argumenter en faveur d’un système éventuellement moins précis dans l’absolu, mais qui répond mieux aux attentes éthiques (et aux besoins) des utilisateurs.

V Conclusion

{comment}

In this article, we have seen that the solutions for solving human value conflicts encountered in the presence of an AI system are never carried out only by IT specialists. The solution can emerged from the exchange between the various stakeholders.

innovations based on artificial intelligence techniques appear to be a clever way of answering questions that have fundamental repercussions on society and risk undermining certain human values. At the time of writing, we are in an unprecedented situation of home lockdown due to the COVID-19 crisis. The 2019-2020 coronavirus pandemic has spread all over the Earth, and this major international public health problem is causing serious socio-economic disruption in the world, including the greatest global recession in nearly a century.

The world is undeniably changing, and this health-related event can be seen as a trigger for the extreme reactions of people, with the appearance of behaviors of great solidarity between individuals and assistance to the most deprived, as well as the moment when selfishness reactions and feelings of xenophobia are exacerbated (e.g., against people perceived as being Chinese because they are seen as the source of the infection).

AI is often associated with robotization, and robotization can be understood as “automation” but also as “the process of turning a human being into a robot.” That is why the arrival of AI can scare some individuals, giving rise to feelings of technophobia with the fear of mass unemployment due to jobs performed by robots or intelligent computer systems. Moreover, there is a rejection of AI because this technology may be seen as a kind of automated and not human –if not “inhuman”– thinking. The species of the human being is Homo sapiens, which means “wise man” in Latin. The human being is not simply characterized as being intelligent, that is to say having an ability to find solutions to complex problems, but also has a nature of a discerning, wise, and sensible being.

The events that we are currently experiencing –COVID-19 crisis, climate change– produce abrupt changes in all areas, bringing new challenges in the fields of health, economy, ecology, security, justice, science, culture…With the collective intelligence of human beings, let AI and ICT be the means pushing us to be innovative in order to respond to these problems in an appropriate manner, and may this response be done wisely in order to guarantee respect for human values.

{comment}

SPECIAL SESSION ON “Ethics in AI, Cyber-Physical and IoT Systems” under the frame of IEEE SMC 2020, October 11-14, 2020 - Toronto, Canada.

V-a Session description

At present time, some people are scared of recent technological developments in the field of Systems, Man and Cybernetics. Most concerns are based on the use of critical data with or without the consent of the user(s). On the other hand, intelligent, self-deciding and proactive systems can also be perceived as a menace. To address these issues and safely coexist in harmony, the deployment of resources interacting with or deciding about human beings must be ethical.

The scope of this Special Session considers the role of ethics in the design of computer systems in the domain of Internet of Things (IoT) and/or Cyber-Physical Systems (CPS) in which Artificial Intelligence (AI) is an important component. Indeed, taking ethics into account requires designing specific models around which to build the systems. Such models can be built a priori or result from an analysis of the behavior of existing systems. Applications can been found in particular healthcare and social services, but also in industries and services.

The topics of interest of this special session include, but are not limited to:

  • Designing models of ethical behavior

  • Implementing models of ethical behavior

  • Interpreting Ethics models

  • Analyzing and Controlling ethical behavior

  • Detecting unethical behavior in cyber-physical systems

  • Combination of Ethics and Knowledge management

  • Inferring Ethics models from IoT/CPS data

  • Ethical behavior in groups of cognitive agents

  • Ethics Simulation in Human-CPS Interaction

  • Ethics Simulation in AI-based IoT Objects

  • Ontologies for Ethics, Ethics models

  • Data protection

  • Ethical data-driven solutions

  • Ethical decision-making by machines

  • Trust and Ethics

This special session is a result of a join effort among the French, Italian, Portuguese and Spanish SMC Chapters.

V-B Submission

Papers must be submitted electronically for peer review through PaperCept by April 14, 2020: http://controls.papercept.net/conferences/scripts/start.pl. In PaperCept, click on the SMC 2020 link “Submit a Contribution to SMC’20” and follow the steps. The code of the special session is: 3v4g3

All papers must be written in English and should describe original work. The length of the paper is limited to a maximum of 6 pages (in the standard IEEE conference double column format).

Footnotes

  1. https://futureoflife.org/ai-principles/
  2. https://www.montrealdeclaration-responsibleai.com/
  3. https://www.aiforhumanity.fr/en/
  4. 1Dlab — Innovation, Culture & Territories: http://en.1d-lab.eu/
  5. Patent FR3046269, inpi, June 2017
  6. ISTEX — Excellence Initiative in Scientific and Technical Information: https://www.istex.fr/
  7. ISTEX — Excellence Initiative in Scientific and Technical Information: https://www.istex.fr/

References

  1. H. T. Al-Natsheh, L. Martinet, F. Muhlenbach, F. Rico and D. A. Zighed (2017) Semantic search-by-examples for scientific topic corpus expansion in digital libraries. In 2017 IEEE International Conference on Data Mining Workshops, ICDM Workshops 2017, New Orleans, LA, USA, November 18-21, 2017, pp. 747–756. External Links: Link Cited by: §IV, §IV.
  2. H. T. Al-Natsheh, L. Martinet, F. Muhlenbach, F. Rico and D. A. Zighed Metadata enrichment of multi-disciplinary digital library: A semantic-based approach. In Digital Libraries for Open Knowledge, 22nd International Conference on Theory and Practice of Digital Libraries, TPDL 2018, Porto, Portugal, September 10-13, 2018, Cited by: §IV.
  3. J. Angwin, J. Larson, S. Mattu and L. Kirchner (2016)(Website) External Links: Link Cited by: §I.
  4. J. Beel, B. Gipp, S. Langer and C. Breitinger (2016) Research-paper recommender systems: a literature survey. International Journal on Digital Libraries 17 (4), pp. 305–338. External Links: Link Cited by: §IV, §IV.
  5. O. Berger-Tal, J. Nathan, E. Meron and D. Saltz (2014) The exploration-exploitation dilemma: a multidisciplinary framework. PLoS ONE 9 (4: e95693). External Links: Link Cited by: §IV.
  6. J. Bonnefon, A. Shariff and I. Rahwan (2016-06) The social dilemma of autonomous vehicles. Science 352 (6293), pp. 1573–1576. External Links: Link, Document Cited by: §II-B.
  7. M. P. Coelho and J. Z. Mendes (2019-08) Digital music and the “death of the long tail”. Journal of Business Research 101, pp. 454–460. External Links: Link Cited by: §III, §III.
  8. J. Dastin (2018)(Website) External Links: Link Cited by: §I.
  9. J. Deonna and E. Tieffenbach (Eds.) (2018) Petit traité des valeurs. Editions d’Ithaque. Cited by: §II-B.
  10. V. Dignum (2019) Responsible artificial intelligence: how to develop and use ai in a responsible way. Springer. Cited by: §II-B.
  11. B. Friedman, D. G. Hendry and A. Borning (2017) A survey of value sensitive design methods. Foundations and Trends in Human-Computer Interaction 11 (2), pp. 63–125. External Links: Link Cited by: §II-C.
  12. B. Friedman and D. G. Hendry (2019-05) Value Sensitive Design: Shaping Technology with Moral Imagination. The MIT Press. External Links: ISBN 9780262351690, Document, Link Cited by: §II-C.
  13. B. Friedman, P. H. Kahn Jr. and A. Borning (2013) Value sensitive design and information systems. In Early engagement and new technologies: Opening up the laboratory, N. Doorn, D. Schuurbiers, I. van de Poel and M. E. Gorman (Eds.), Dordrecht, pp. 55–95. External Links: Link Cited by: §II-B, §II-C.
  14. D. Klitou (2014) Privacy-invading technologies and privacy by design: safeguarding privacy, liberty and security in the 21st century. T.M.C. Asser Press. Cited by: §II-A.
  15. T. Kohonen (2001) Self-organizing maps. Springer. Cited by: §III.
  16. P. W. Langley, H. A. Simon, G. Bradshaw and J. M. Zytkow (1987) Scientific discovery: computational explorations of the creative process. The MIT Press, Cambridge, MA. Cited by: §IV.
  17. P. -. Lhérisson, F. Muhlenbach and P. Maret (2017) Fair recommendations through diversity promotion. In Advanced Data Mining and Applications - 13th International Conference, ADMA 2017, Singapore, November 5-6, 2017, Proceedings, pp. 89–103. External Links: Link, Document Cited by: §III.
  18. P. Lhérisson, F. Muhlenbach and P. Maret Application mobile pour l’évaluation d’un algorithme de calcul de distance entre des items musicaux. In 17ème Journées Francophones Extraction et Gestion des Connaissances, EGC 2017, 24-27 Janvier 2017, Grenoble, France, F. L. Gandon and G. Bisson (Eds.), Cited by: §III.
  19. G. Marchionini (2006) Exploratory search: from finding to understanding. Communications of the ACM (CACM) 49, pp. 41–46. External Links: Link Cited by: §IV.
  20. B. Mepham, M. Kaiser, E. Thorstensen, S. Tomkins and K. Millar (2006-02) Ethical matrix manual. Technical report LEI, The Hague. Cited by: §II-B.
  21. B. Mepham (2000) A framework for the ethical analysis of novel foods: the ethical matrix. Journal of Agricultural and Environmental Ethics 12, Issue 2, pp. 165–176. Cited by: §II-B.
  22. T. Mikolov, I. Sutskever, K. Chen, G. Corrado and J. Dean (2013) Distributed representations of words and phrases and their compositionality. CoRR abs/1310.4546. Cited by: §IV, §IV.
  23. Montréal Declaration on Responsible AI (2018)(Website) External Links: Link Cited by: §II-B.
  24. F. Muhlenbach and I. Sayn (2019) Artificial intelligence and law: what do people really want? Example of a French multidisciplinary working group. In Proceedings of the Seventeenth International Conference on Artificial Intelligence and Law, ICAIL 2019, Montreal, QC, Canada, June 17-21, 2019, pp. 224–228. External Links: Link, Document Cited by: §II-A.
  25. C. O’Neil (2016) Weapons of math destruction: how big data increases inequality and threatens democracy. Crown Publishers. Cited by: §I.
  26. E. Pariser (2011) The filter bubble: what the internet is hiding from you. Penguin Press, New York. Cited by: §I, §III, §IV.
  27. K. R. Popper (2002) The logic of scientific discovery. Routledge Classics, London and New York. Cited by: §IV, §IV.
  28. J. Rose, J. S. Persson, L. T. Heeager and Z. Irani (2015) Managing e-government: value positions and relationships. Information Systems Journal 25 (5), pp. 531–571. External Links: Link, Document Cited by: §II-B.
  29. S. J. Russell and P. Norvig (2016) Artificial intelligence –a modern approach. Third edition, Pearson Education Limited. Cited by: §I.
  30. S. Russell (2016-05) Should we fear supersmart robots?. Scientific American 314, pp. 58–59. External Links: Document Cited by: §II-B.
  31. M. Scheler (1973) Formalism in ethics and non-formal ethics of values. Northwestern University Press. Cited by: §II-B.
  32. S. H. Schwartz (1992) Universals in the content and structure of values: theoretical advances and empirical tests in 20 countries. In Advances in Experimental Social Psychology, M. P. Zanna (Ed.), Vol. 25, pp. 1–65. External Links: Link Cited by: §II-B.
  33. Scientific and Technical Information Department – CNRS (2016) White paper – open science in a digital republic. OpenEdition Press, Marseille. External Links: Document, Link Cited by: §IV, §IV.
  34. Scientific and Technical Information Department – CNRS (2017) White paper – open science in a digital republic – strategic guide. OpenEdition Press, Marseille. External Links: Document, Link Cited by: §IV, §IV.
  35. T. J. Sejnowski (2018) The deep learning revolution. The MIT Press. Cited by: §I.
  36. J. van den Hoven, P. E. Vermaas and I. van de Poel (Eds.) (2015) Handbook of ethics, values, and technological design: sources, theory, values and application domains. Springer Netherlands. External Links: Link Cited by: §II-C, §II-C.
  37. C. Wylie (2019) Mindf*ck: cambridge analytica and the plot to break america. Penguin Random House LLC., New York. Cited by: §I.
Comments 0
Request Comment
You are adding the first comment!
How to quickly get a good reply:
  • Give credit where it’s due by listing out the positive aspects of a paper before getting into which changes should be made.
  • Be specific in your critique, and provide supporting evidence with appropriate references to substantiate general statements.
  • Your comment should inspire ideas to flow and help the author improves the paper.

The better we are at sharing our knowledge with each other, the faster we move forward.
""
The feedback must be of minimum 40 characters and the title a minimum of 5 characters
   
Add comment
Cancel
Loading ...
417037
This is a comment super asjknd jkasnjk adsnkj
Upvote
Downvote
""
The feedback must be of minumum 40 characters
The feedback must be of minumum 40 characters
Submit
Cancel

You are asking your first question!
How to quickly get a good answer:
  • Keep your question short and to the point
  • Check for grammar or spelling errors.
  • Phrase it like a question
Test
Test description