The paper is a systematic literature review of the emerging field of AI4SG and relational AI ethics. Since these two fields have not interacted with each other, the paper aims to lay the foundation for such an interface. The paper argues that such an interface is exigent because AI4SG is emerging as a normative field of AI ethics. Interfacing with existing literature on relational ethics can help interrogate the normative core of AI4SG and open up both the 'social' and the 'technical' as more than instrumental concerns. AI4SG, AI ethics, relational, systematic literature review, technology *Corresponding author. †These authors contributed equally.
The paper aims to bring into conversation two fields of AI ethics through a systematic literature review of the two fields, that of AI4SG and the relational ethics approach to AI.
The gold rush in the last few years concerning ethical AI [
While ethical debates in the field of AI ethics have flourished in recent years at a growing speed, the
field of AI4SG is slowly emerging as a justificatory ground for developing AI. As the paper notes in a
later section, the AI4SG discourse has been funded and facilitated by powerful actors such as United
Nations, international non-governmental organizations, consultancy companies and big tech. Against
this backdrop, critical interrogation of the very notion of AI4SG is exigent. While AI4SG privileges
normative ethical approaches, the field of AI ethics on the other hand, while moving towards
convergences, is still open to contestation, with relational ethics offering the most radical critique to
normative ethical approaches that posit a transcendentalist and/or vertical systems of morality and fall
back on a problem/solution framework. Relational ethical approaches, on the other hand, pivot around
ontological heterogenesis [14], horizontal relations, and situated and partial perspectives [50] in which
moral systems are emergent and processual [
2023 Copyright for this paper by its authors. CEUR
ceur-ws.org ISSN1613-0073
It is in this light that the paper does a systematic literature review of AI4SG and relational AI ethics to understand how and if at all they interface. The key finding of the paper is that there has been no interface between these two fields in major peer-reviewed publications. This review aims to lay the ground for the future interface between the two domains.
Following the introduction, a brief review of existing surveys is provided, and the methodology of the paper is elaborated. The fourth section presents a review and commentary on the AI4SG discourse and the fifth section reviews the field of relational AI ethics. A brief conclusion, limitations, and need for future work are noted.
Previous surveys of AI4SG, while larger in scope, did not conduct a systematic literature review of the
field of AI4SG [
We depart from such approaches for a very simple reason. AI4SG is not a category or approach
which can be given a hall pass in the debates of AI ethics because of their claims of ‘social good’.
Instead, unpacking the very category of ‘social’ and ‘good’ is necessary. It is important to note that in
the discipline of sociology, the very category of ‘social’ is contestable and no single consensus exists.
In recent years, the ‘social’ has emerged as a domain with a multiplicity of connections that spans
human and non-human domains [64]. This fundamentally open view of the ‘social’ means that political
consensus remains a myth that perpetuates hegemonic discourses and ideas [
On the other hand, while relational ethics has emerged in certain subdomains of AI ethics literature more strongly than others, such as in the case of robot ethics and AI in medical domains, relational ethical approaches remain relatively marginal because they are difficult to operationalize within solutionist fields such as engineering, policy-making and governance. At the time of writing, we didn’t find any comprehensive review of the field of relational AI ethics.
To conduct the systematic literature review, we referred to Scopus and Web of Science. Limiting the review to these two indexes has definite shortcomings as they are heavily skewed towards English language publications and don’t include a wide variety of open access, pre-print and experimental publications. Yet, limiting ourselves to these two indexes also allows us to map the state of the field in relatively high-impact journals and conference proceedings.
The paper also refers to certain developments in broad fields of social sciences to establish links and conversations which would otherwise not be possible through a traditional systematic literature review. For example, while AI4SG invokes the concept of ‘social’, the discourse on AI4SG has largely bypassed discussions on ‘social’ that have transpired over the past two decades in sociology. As anthropologists/sociologists reviewing the field of AI4SG, we cannot help but bring these blatantly neglected discussions to light. The discourse of AI4SG also invokes uncritically the United Nations (UN) Sustainable Development Goals (SDGs). Hence, to situate such arbitrary external relationships it was necessary to situate such invocations alongside pre-existing critiques of such frameworks to provide the reader with a balanced view of a relatively nascent field.
For AI4SG, two keywords ‘({artificial intelligence for social good} OR {artificial intelligence for
good} OR {AI4SG} OR {AI4G})’ were queried (see Figure 1 for PRISMA flow [
For the second part of the review, the search phrase ‘{AI OR {artificial AND intelligence}} AND
relational AND ethics’ was used to spread the net as wide as possible (see Figure 2 for PRISMA flow
and Table 2 for an overview of the documents). While the field of AI ethics has grown over the past
few years, relational approaches in AI ethics are still a relatively small field with higher activity in the
domains of health and robot rights and ethics. Five important publications in the field that the index
searches did not throw up were reviewed manually and are referred to in the narrative of the paper for
their significant relevance to the field of relational ethics and artificial intelligence [
n o i t a c i f i t n e d I g n i n e e r c S d e d u l c n I Records screened (n = 56) Records excluded (n = 16)
screening: Duplicate records removed (n = 13)
automation tools (n = 0)
AI4SG (and other ancillary notions such as AI4G) has emerged in the past few years to distinguish the
usage of AI from “malicious use, e.g. DeepNude, or for highly dubious purposes” [91:3]. One of the
earliest celebrations of AI4SG was in a workshop report published by Hager et al. [
Though Hager et al. [
Cowls et al. [25:112] define AI4SG as: “the design, development and deployment of AI systems in ways that help to (i) prevent, mitigate and/or resolve problems adversely affecting human life and/or the wellbeing of the natural world, and/or (ii) enable socially preferable or environmentally sustainable developments, while (iii) not introducing new forms of harm and/or amplifying existing disparities and inequities.”
While such a definition of AI4SG is deceptively simple, it bypasses and in turn obfuscates the political nature of the existing problems. This obfuscation is particularly evident in Cowls et al. [25] as they deploy the logic which announces that “AI4SG = AI×SDGs” [25:112]. Though they note the limitations of such an equation which links AI4SG to the UN Sustainable Development Goals they argue for such an approach under the assumption that “SDGs offer clear, well defined and shareable boundaries to identify positively what is socially good AI” and that “SDGs are internationally agreed on goals for development and have begun informing relevant policies worldwide, so they raise fewer questions about relativity and cultural dependency of values” [25:112]. Such framing has some glaring shortcomings that Cowls et al. choose to ignore. UN SDGs, like most United Nations consensus exercises, are not legally binding nor represent goals and principles that political groups intra-nationally or trans-nationally agree upon. Secondly, SDGs, as Telleria and Garcia-Arias [95:15] argue, is an “empty signifier that keeps disparate and even contradictory demands united”. It does this by constructing “a fantasmatic explanation of international development and sustainability issues that conceals the antagonistic dimension of social, political and economic issues” [95:15]. In a later special issue introduction edited by Cowls, Cowls notes contrary to such a straightforward and simple solution that “The contributions make clear that neither ‘AI’ nor ‘social good’ should be thought of as uncontested or incontestable terms, and we should remain wary of the twin dangers of unjustified hype and unseen harm arising from the continued growth of interest in, and application of, AI” [24:54].
Cowls et al. [25] accompanied by Floridi et al. [
4.2.
While the field of AI ethics and its ancillary fields such as robot ethics have flourished over the last
decade with a range of viewpoints, AI4SG has emerged as a normative ethical field. Following Cowls
et al. [25] and Floridi et al. [
Another interesting phenomenon is the presence of the AI4SG or AI4G terms in the keywords,
funding details or as a stray reference in the abstract with no follow-through in the paper
[
The influence of normative ethics doesn’t simply end at an explicit link between AI4SG and the UN
SDGs. Mabaso [
A contrarian attitude is visible in a minor domain of AI4SG which is concerned with thinking through adversarial strategies such as the inclusion of “well-calibrated noise, imperceptible to humans” [98:1] in datasets to reduce the risk of genomic data being matched to images of human faces. Interestingly, while [98] doesn’t use the phrase AI4SG anywhere in the article, one of the project titles of a National Science Foundation-funded project referred explicitly to “Adversarial artificial intelligence for social good” [98:9].
4.3.
A slow but steady body of work is today emerging from sciences, engineering and other applied
disciplines which now deploy the notion of AI4SG to justify and frame their research work
[
While addressing primarily AI4SG discourse within the domain of health, Holzmeyer [56] levies a wide-ranging set of critiques against AI4SG which largely mirrors various critiques of AI. Holzmeyer anchors their critique of AI4SG around four axes, distraction from root causes, vulnerabilities and risks of big data, politics of data and expertise that frames computational solutions as better and perpetuation of “cycles of (re)validating that environmental and social variables matter to health” leading to “continued rationalization of inaction” [56:109]. A broader radical critique that Holzmeyer levies are grounded on the fact that AI4SG is “led by many of the same corporate actors that are incubating AI systems broadly” in collaboration with hegemonic actors such as the Big Tech, United Nations, established NGOs and consultancies such as McKinsey, Deloitte, Accenture. The involvement of such powerful and hegemonic actors in the AI4SG discourse necessitates a radical political interrogation and is proof that AI4SG is not a mere technical problem to be addressed via mathematical formulas and code but necessitates a larger socio-technical view of the situation. Various actors involved in AI4SG and AI are limited by “their inattention to or blindspots around larger social, political and economic systems often hamper consideration of the fuller range of social justice issues at stake” [56:112].
Madianou’s [
Sapignoli stresses the increasing usage of AI in global governance and human rights domains
arguing for ethnographic approaches to study sociotechnical systems calling for an
anthropologyinflected “AI-turn in international governance” which “would be an ethics of realism that takes into
consideration and reveals the cultural, political, and economic context in which AI programs are
embedded and how their applications are translated in diverse cultural contexts and jurisdictions with
different consequences” [83:298]. Sapignoli’s call stresses on both the theoretical—“will not only
generate new insights about the nature of knowledge production and decision-making, but could also
bring about a transformation of anthropological theory”—and the applied dimension—“Just as social
scientists have engaged in a critical analysis of the rule of law, these scientists should unpack the rule
of AI design, and hopefully mitigate its consequences, both intended and unintended” [83:298].
While critiques of AI4SG have grown, so have interfaces with various other theoretical, philosophical
and methodological approaches. The call for more meaningful stakeholder participation has emerged
as one of the calls for operationalizing AI4SG [63]. A more nuanced elaboration of stakeholder
consultation is elaborated by Bondi et al. [13] who argue for an interfacing of the capabilities approach
to operationalise AI4SG. The cost-benefit analysis in such an instance is not tied to global SDG goals
but is evaluated by considering the capabilities of the various participants noting that “the goals of the
project are properly aligned to do so without negatively impacting other capabilities” [13:7]. Berberich
et al. [
Schiff et al. [
The existing literature presents relational ethics as an approach that proposes a "fundamental shift—
from rational to relational—in thinking about personhood, data, justice and everything in between"
[11:1]. It hopes to move away from normative ethical principles that are "often formulated as
problem/solution" [11:1] towards a more critical evaluation of the field that calls for "rethinking of
justice and ethics as a set of broad, contingent, and fluid concepts and down-to-earth practices that are
best viewed as a habit and not a mere methodology of data science" [11:1]. One of the earliest instances,
where the power of a relativist stance is vindicated is in a 1998 article by Rössler & Matsuno [
This earliest distinction is interesting as these fault lines can be traced in the literature even today. While not mutually exclusive, the literature on AI relational ethics is divided along two axes: First axis concentrates on the ‘what is’ question, which debunks the property-based approach to the question of bots and their moral status in favour of relational ontology wherein “The partners do not pre-exist their relating; the partners are precisely what comes out of the inter- and intra-relating of fleshly, significant, semioticmaterial being’’ (Haraway 2008, p. 165 as cited in [22:10]); second axis addresses the ‘how’ part of relational ethics which foregrounds ethical approaches/principles that go beyond problem/solution dyad to address power imbalances, inequalities, and relations of injustices inherent in the design and development of new and existing technologies.
While the former hopes to completely redefine the notion of the self and the other when it interrogates the moral status of AI and its relationship to the human—a deep technology view to the question of human, machine or animal—the latter hopes to arrive at relational yet standard approaches to design, develop and evaluate AI systems that minimise algorithmic harms and injustices. Both these axes, however, oppose other normative ethical frameworks that maintain “a strict divide between the creator (spiritual) and the created (non-spiritual)… with humans in the position of creator” [18:57].
These critiques point towards the limits of principalism and approaches to moral standing that assume an “ontological platform onto which morality is mounted” [22:9]. Relational ethics, on the other hand, argues that, “the moral-epistemic ground…is not situated in the ‘essence’ of the [object], but is something that comes into being, is gradually revealed and constructed as we get to know objects inrelation, as human beings who interact with them, watch them, call them by name, etc.— whether as scientists, philosophers, or as lay persons” [22:16]. 5.2.
The first axis that is invested in thinking through the deep technology problematic of relational ethics has been gaining slow yet steady ground since the 2010s focusing on theorizing human-bot relations while intervening in the debate on the moral status and rights of robots and AI. Coeckelbergh [18] opened the field by interrogating the anthropocentric bias in thinking about the “technological future” arguing in favour of a “deep relational view on human being and self” that borrowed from ecological and eastern philosophical approaches. Cockelbergh and Gunkel have been the most prominent voices foregrounding a relational view of the question of the moral status of bots and human-bot relations. They reframe the standard ethical question “does the animal have a face” to ask “[W]hat does it take for an animal to supervene and be revealed as having face in the Levinasian sense?” which demands that we learn how to respond and what responding will mean in such a case [22:9]. Thus, in asking the question of response-ability, they replace the figure of a moral philosopher and observer with an active practitioner and sense-maker who is part of a lifeworld that operates on a certain ‘we’.
The question of relations, of how to respond to plants and other non-humans, thus is a political
societal and collective question [20]. A property-based approach to the question of the moral/legal status
of others privileges ontological properties such as consciousness, sentience, intelligence, suffering etc
as valid grounds to decide an object’s moral/legal status. Developing his critique of this approach
further, Gunkel[
The second axis moves away from the question of robot rights and moral/legal status to operationalize relational ethics more concretely. Birhane makes a clear attempt to call for relational approaches to AI ethics, harms and injustice to make a case for reconceptualizing the notion of “personhood, data, justice and everything in between” [11:1]. Birhane responds to the growing trend in the field that approaches AI ethics as a technical problem, and the subsequent flourishing of AI ethics toolkits, to argue for an understanding of AI ethics that goes “beyond technical solutions” [11:1].
While Birhane and others limit themselves to a critical examination of the field rather than offering solutions, a significant effort has been made, both before and after Birhane, to arrive at relational reconceptualizations of specific ethical principles such as transparency, responsibility, fairness, intelligence as well as the need to understand AI as an algorithmic assemblage with deeply relational affordances [29]. Ananny [3:7] puts forth a sociotechnical definition of AI systems, as a unit of AI ethical analysis, that they call networked information algorithm (NIA), “as an assemblage …of institutionally situated computational code, human practices, and normative logics that creates, sustains, and signifies relationships among people and data through minimally observable, semiautonomous action. Although code, practices, and norms may be observed individually in other contexts, their full ‘‘meaning and force ... can only be understood in terms of relations with other modular units…’’.
Ananny [3:2] identifies three dimensions for scrutinizing AI ethics and holding algorithmic assemblages accountable: “The ability to convene people by inferring associations from computational data, the power to judge similarity and suggest probable actions, and the capacity to organize time and influence when action happens”
They too distinguish their approach from “mathematical, mechanistic focus” [3:5] which only asks if a code is biased or not to instead asks whether different assemblages ‘‘help us get into satisfactory relation with other parts of our experience’’ [3:7].
Focussing on the use of AI in education, Henry & Oliver [52] make a similar move to argue that “ethics should not be understood as abstract values or design decisions, but as socio-technical achievements” [52:330]. They deploy Puig de la Bellasca’s speculative critique to offer a vision of caring that is ‘an analytics or provocation’ that presupposes heterogeneity as the ontological ground. In exploring the tension between ethics and justice, they foreground questions of not only “for whom” but who cares, what for and why do we care?” [52:334].
Along similar lines, others have focussed on identifying what new demands AI makes of ethical
frameworks and how these frameworks and principles might be reconceptualized in light of these
demands. Coeckelbergh focuses on the question “of responsibility attribution for artificial intelligence
technologies” [21:2052] and introduces another actor in the responsibility relation: moral patients, who
are affected by the action of the agent and those to whom moral agents are responsible and answerable.
Giovanola and Tiribelli [
Others in the field have focussed on identifying the social impacts of AI systems in domains such as
healthcare, education, warfare, and labour to ask for deeper stakeholder engagement to balance
considerations of relational care, safety, privacy and transparency [53:582]; Döbler & Bartnik [29]
focus on technological affordances that can reveal, manipulate and conceal possible actions. Schoenherr
[
Papers
[
Relational ethics is often critiqued as a metaphysical approach that says more about the nature of being
without generating a clear ethics i.e., the challenge of deriving an ‘ought’ from an ‘is’ [
Gibert & Martin [
Gordon & Gunkel [
This is an important point to consider and is part of the challenge of a relational approach. A
relational approach to AI ethics borrows from poststructuralist approaches that primarily operate at the
level of critique which can help identify limits of existing knowledge paradigms, ethical frameworks or
standard operations. Recognizing the critical gesture inherent in the relational approach to AI ethics,
Birhane [
As noted, AI4SG is slowly emerging as a normative ethical field within the larger domain of AI ethics with minor contrarian or non-normative approaches. This necessitates interrogating and interfacing non-normative ethical approaches with AI4SG. The dominant approach of linking AI4SG with SDGs doesn’t guard against AI harms and injustices but instead reduces the problem of ‘social’ to a technical process of identifying socially relevant AI systems as identified by the international bureaucratic machine.
Relational ethics take AI assemblage as its unit of analysis that opens up the ‘social’ to the political,
societal and collective question that addresses the relations of power, inequality, harms and injustices
as more than instrumental concerns. To interface AI4SG with relational ethics is to open up the black
box of not just technical concerns but also social relations and see AI ethics as a socio-technical
achievement [
The fundamental limitation of the paper is the usage of mainstream indexes such as Scopus and Web of Science and the inclusion criteria of the English language. Future research in all languages across other indexes could help shed light on how and if at all AI4SG and relational ethics are interfacing in the existing literature.
Building on Birhane’s [
This material is based upon work supported in whole or in part by The Notre Dame-IBM Tech Ethics Lab. Such support does not constitute an endorsement by the sponsor of the views expressed in this publication. [13] Elizabeth Bondi, Lily Xu, Diana Acosta-Navas, and Jackson A. Killian. 2021. Envisioning Communities: A Participatory Approach Towards AI for Social Good. In Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society (AIES ’21), Association for Computing Machinery, New York, NY, USA, 425–436. DOI:https://doi.org/10.1145/3461702.3462612 [14] Rosi Braidotti. 2006. Transpositions: on nomadic ethics. Polity Press, Cambridge, UK ; Malden,
MA. [15] Morgan Briggs. 2021. SatDash: An Interactive Dashboard for Assessing Land Damage in Nigeria and Mali. In ACM SIGCAS Conference on Computing and Sustainable Societies (COMPASS ’21), Association for Computing Machinery, New York, NY, USA, 100–114.
DOI:https://doi.org/10.1145/3460112.3471949 [16] Marianna Capasso and Steven Umbrello. 2022. Responsible nudging for social good: new healthcare skills for AI-driven digital personal assistants. Med Health Care and Philos 25, 1 (March 2022), 11–22. DOI:https://doi.org/10.1007/s11019-021-10062-z [17] William Clapton and Laura J. Shepherd. 2019. Ethics Ex Machina : popular culture and the plural futures of politics. Australian Journal of Political Science 54, 4 (October 2019), 531–542.
DOI:https://doi.org/10.1080/10361146.2019.1663400 [18] Mark Coeckelbergh. 2013. Pervasion of what? Techno–human ecologies and their ubiquitous spirits. AI & Soc 28, 1 (February 2013), 55–63. DOI:https://doi.org/10.1007/s00146-012-0418-y [19] Mark Coeckelbergh. 2016. Responsibility and the Moral Phenomenology of Using Self-Driving Cars. Applied Artificial Intelligence 30, 8 (September 2016), 748–757.
DOI:https://doi.org/10.1080/08839514.2016.1229759 [20] Mark Coeckelbergh. 2018. What do we mean by a relational ethics?: Growing a relational approach to the moral standing of plants, robots, and other non-humans. In Plant Ethics, Angela Kallhoff, Marcello Di Paola and Maria Schörgenhumer (eds.). Routledge. [21] Mark Coeckelbergh. 2020. Artificial Intelligence, Responsibility Attribution, and a Relational Justification of Explainability. Sci Eng Ethics 26, 4 (August 2020), 2051–2068.
DOI:https://doi.org/10.1007/s11948-019-00146-8 [22] Mark Coeckelbergh and David J. Gunkel. 2014. Facing Animals: A Relational, Other-Oriented Approach to Moral Standing. J Agric Environ Ethics 27, 5 (October 2014), 715–733.
DOI:https://doi.org/10.1007/s10806-013-9486-3 [23] Marta R. Costa-jussà, Esther González, Asuncion Moreno, and Eudald Cumalat. 2020. Abusive language in Spanish children and young teenager’s conversations: data preparation and short text classification with contextual word embeddings. In Proceedings of the Twelfth Language Resources and Evaluation Conference, European Language Resources Association, Marseille, France, 1533–1537. Retrieved June 2, 2023 from https://aclanthology.org/2020.lrec-1.191 [24] Josh Cowls. 2021. ‘AI for Social Good’: Whose Good and Who’s Good? Introduction to the Special Issue on Artificial Intelligence for Social Good. Philos. Technol. 34, 1 (November 2021), 1–5. DOI:https://doi.org/10.1007/s13347-021-00466-3 [25] Josh Cowls, Andreas Tsamados, Mariarosaria Taddeo, and Luciano Floridi. 2021. A definition, benchmark and database of AI for social good initiatives. Nat Mach Intell 3, 2 (February 2021), 111–115. DOI:https://doi.org/10.1038/s42256-021-00296-0 [26] Veronica Dahl and Juan José Moreno-Navarro. 2022. Doughnut Computing in City Planning for Achieving Human and Planetary Rights. In Bio-inspired Systems and Applications: from Robotics to Ambient Intelligence (Lecture Notes in Computer Science), Springer International Publishing, Cham, 562–572. DOI:https://doi.org/10.1007/978-3-031-06527-9_56 [27] Giulia De Togni, Sonja Erikainen, Sarah Chan, and Sarah Cunningham-Burley. 2021. What makes AI ‘intelligent’ and ‘caring’? Exploring affect and relationality across three sites of intelligence and care. Social Science & Medicine 277, (May 2021), 113874.
DOI:https://doi.org/10.1016/j.socscimed.2021.113874 [28] Nicole Dewandre. 2019. Humans as relational selves. AI & Soc 34, 1 (March 2019), 95–98.
DOI:https://doi.org/10.1007/s00146-017-0700-0 [29] Niklas Döbler and Clemens Bartnik. 2022. Normative Affordances Through and By Technology: Technological Mediation and Human Enhancement. International Journal of Interactive Multimedia and Artificial Intelligence 7, (September 2022), 14–23. DOI:https://doi.org/10.9781/ijimai.2022.09.006 in the diagnosis of rare disease. BMC medical ethics 23, 1 (November 2022).
DOI:https://doi.org/10.1186/s12910-022-00842-4 [49] Jeffrey T Hancock, Mor Naaman, and Karen Levy. 2020. AI-Mediated Communication: Definition, Research Agenda, and Ethical Considerations. Journal of Computer-Mediated Communication 25, 1 (March 2020), 89–100. DOI:https://doi.org/10.1093/jcmc/zmz022 [50] Donna Haraway. 1988. Situated Knowledges: The Science Question in Feminism and the Privilege of Partial Perspective. Feminist Studies 14, 3 (1988), 575–599.
DOI:https://doi.org/10.2307/3178066 [51] Jamie Harris and Jacy Anthis. 2021. The Moral Consideration of Artificial Entities: A Literature Review. Science and Engineering Ethics 27, (August 2021). DOI:https://doi.org/10.1007/s11948021-00331-8 [52] Jade Vu Henry and Martin Oliver. 2022. Who Will Watch the Watchmen? The Ethico-political Arrangements of Algorithmic Proctoring for Academic Integrity. Postdigit Sci Educ 4, 2 (April 2022), 330–353. DOI:https://doi.org/10.1007/s42438-021-00273-1 [53] Anita Ho. 2020. Are we ready for artificial intelligence health monitoring in elder care? BMC
Geriatr 20, 1 (December 2020), 358. DOI:https://doi.org/10.1186/s12877-020-01764-9 [54] C. W. L. Ho, D. Soon, K. Caals, and J. Kapur. 2019. Governance of automated image analysis and artificial intelligence analytics in healthcare. Clin Radiol 74, 5 (May 2019), 329–337.
DOI:https://doi.org/10.1016/j.crad.2019.02.005 [55] Andreas Holzinger, Anna Saranti, Alessa Angerschmid, Carl Orge Retzlaff, Andreas Gronauer, Vladimir Pejakovic, Francisco Medel-Jimenez, Theresa Krexner, Christoph Gollob, and Karl Stampfer. 2022. Digital Transformation in Smart Farm and Forest Operations Needs HumanCentered AI: Challenges and Future Directions. Sensors 22, 8 (January 2022), 3043.
DOI:https://doi.org/10.3390/s22083043 [56] Cheryl Holzmeyer. 2021. Beyond ‘AI for Social Good’ (AI4SG): social transformations—not tech-fixes—for health equity. Interdisciplinary Science Reviews 46, 1–2 (April 2021), 94–125.
DOI:https://doi.org/10.1080/03080188.2020.1840221 [57] Meng-Leong How, Sin-Mei Cheah, Aik Cheow Khor, and Yong Jiet Chan. 2020. Artificial Intelligence-Enhanced Predictive Insights for Advancing Financial Inclusion: A Human-Centric AI-Thinking Approach. Big Data and Cognitive Computing 4, 2 (June 2020), 8.
DOI:https://doi.org/10.3390/bdcc4020008 [58] Liza Ireni-Saban and Maya Sherman. 2021. Ethical Governance of Artificial Intelligence in the
Public Sector. Routledge, London. DOI:https://doi.org/10.4324/9781003106678 [59] Lakshmi Shankar Iyer and Sirish Venkatagiri. 2019. Is Digital GOOD for providing water security?- A case study of India. International Journal of Advanced Science and Technology 28, 16 (November 2019), 1074–1092. [60] Nancy S. Jecker. 2021. Can we wrong a robot? AI & Soc (November 2021).
DOI:https://doi.org/10.1007/s00146-021-01278-x [61] Nancy S. Jecker, Caesar A. Atiure, and Martin Odei Ajei. 2022. The Moral Standing of Social Robots: Untapped Insights from Africa. Philos. Technol. 35, 2 (April 2022), 34.
DOI:https://doi.org/10.1007/s13347-022-00531-5 [62] Anna Jobin, Marcello Ienca, and Effy Vayena. 2019. The global landscape of AI ethics guidelines.
Nat Mach Intell 1, 9 (September 2019), 389–399. DOI:https://doi.org/10.1038/s42256-019-00882 [63] Meghana Kshirsagar, Caleb Robinson, Siyu Yang, Shahrzad Gholami, Ivan Klyuzhin, Sumit Mukherjee, Md Nasir, Anthony Ortiz, Felipe Oviedo, Darren Tanner, Anusua Trivedi, Yixi Xu, Ming Zhong, Bistra Dilkina, Rahul Dodhia, and Juan M. Lavista Ferres. 2021. Becoming Good at AI for Good. In Proceedings of the 2021 AAAI/ACM Conference on AI, Ethics, and Society (AIES ’21), Association for Computing Machinery, New York, NY, USA, 664–673.
DOI:https://doi.org/10.1145/3461702.3462599 [64] Bruno Latour. 2005. Reassembling the Social: An Introduction to Actor-Network-Theory. OUP
Oxford. [65] Migle Laukyte. 2019. Against Human Exceptionalism: Environmental Ethics and the Machine Question. In On the Cognitive, Ethical, and Scientific Dimensions of Artificial Intelligence, Don social good: unlocking the opportunity for positive impact. Nat Commun 11, 1 (May 2020), 2468.
DOI:https://doi.org/10.1038/s41467-020-15871-z [97] Sapdo Utomo, A John, Ayush Pratap, Zhi-Sheng Jiang, P. Karthikeyan, and Pao-Ann Hsiung. 2023. AIX Implementation in Image-Based PM2.5 Estimation: Toward an AI Model for Better Understanding. In 2023 15th International Conference on Knowledge and Smart Technology (KST), 1–6. DOI:https://doi.org/10.1109/KST57286.2023.10086917 [98] Rajagopal Venkatesaramani, Bradley A. Malin, and Yevgeniy Vorobeychik. 2021. Reidentification of individuals in genomic datasets using public face images. Science Advances 7, 47 (November 2021), eabg3296. DOI:https://doi.org/10.1126/sciadv.abg3296 [99] Vasiliki Voukelatou, Lorenzo Gabrielli, Ioanna Miliou, Stefano Cresci, Rajesh Sharma, Maurizio Tesconi, and Luca Pappalardo. 2021. Measuring objective and subjective well-being: dimensions and data sources. Int J Data Sci Anal 11, 4 (May 2021), 279–309.
DOI:https://doi.org/10.1007/s41060-020-00224-2 [100] Rosalie Waelen. 2022. Why AI Ethics Is a Critical Theory. Philos. Technol. 35, 1 (February 2022), 9. DOI:https://doi.org/10.1007/s13347-022-00507-5 [101] Hua Wang, Sneha Gupta, Arvind Singhal, Poonam Muttreja, Sanghamitra Singh, Poorva Sharma, and Alice Piterova. 2022. An Artificial Intelligence Chatbot for Young People’s Sexual and Reproductive Health in India (SnehAI): Instrumental Case Study. J Med Internet Res 24, 1 (January 2022), e29969. DOI:https://doi.org/10.2196/29969 [102] Ruitao Xie, Connor Qiu, and Guoping Qiu. 2022. Finding Beautiful and Happy Images for Mental Health and Well-Being Applications. In Pattern Recognition and Computer Vision (Lecture Notes in Computer Science), Springer Nature Switzerland, Cham, 704–717. DOI:https://doi.org/10.1007/978-3-031-18913-5_54