Reflection on Big Data and AI
rz6794
EBSCO-FullText-02_07_2026.pdf
ISSN: 1795-6889
https://ht.csr-pub.eu Volume 20(1), May 2024, 6-24
6
CONCEPTUAL FRAMEWORK FOR ETHICAL ARTIFICIAL INTELLIGENCE DEVELOPMENT IN SOCIAL SERVICES
SECTOR
Miroslavas Seniutis Institute of Sociology and Social Work
Vilnius University
Lithuania
ORCID 0000-0002-8089-3341
Valentas Gružauskas Institute of Computer Science
Vilnius University
Lithuania
ORCID 0000-0002-6997-9275
Angele Lileikiene Lithuania Business College, Lithuania
ORCID 0000-0002-8414-5906
Valentinas Navickas Lithuania Business College, Lithuania
ORCID 0000-0002-7210-4410
Abstract: This research explores the domain of Artificial Intelligence (AI) for social good,
with a particular emphasis on its application in social welfare and service delivery. The
study seeks to establish a universal conceptual framework for ethically integrating AI into
the social services sector, recognizing the sector's significant yet underexplored potential
for AI utilization. The objective is to develop a comprehensive framework applicable to the
ethical deployment of AI in social services, using Lithuania as a case study to illustrate its
practicality. This involves analysing the political discourse on AI, examining its
applications in social welfare, identifying ethical challenges, evaluating the digitalization
progress in Lithuania's public services, and formulating guidelines for AI integration at
various stages of delivering social services. Our methodology is rooted in document
analysis, encompassing a thorough review of both normative and scientific literature
pertinent to the ethical application of AI in social welfare. Key findings reveal that AI's
anticipated positive impacts on diverse social and economic areas, as highlighted in
political declarations, are being partially realized, as corroborated by scientific studies.
Although the global application of AI in social welfare is expanding, Lithuania presents a
unique case with its strategic planning gaps in this sector. The developed conceptual
framework offers vital criteria for the ethical implementation of AI systems designed to be
universally applicable to various stages of social services, accommodating different AI
applications, client groups, and institutional environments.
Keywords: artificial intelligence, social service, ethics, innovations
©2024 Seniutis, Gružauskas, Lileikiene, & Navickas, and the Centre of Sociological
Research, Poland
DOI: https://doi.org/10.14254/1795-6889.2024.20-1.1
This work is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License.
Conceptual Framework for Ethical Artificial Intelligence Development in Social Services Sector
7
INTRODUCTION
Background and Problem Statement
There is a field of research known as Artificial Intelligence1 (AI) for Social Good2, which
focuses on using AI to benefit society as a whole. It often addresses issues that have a broader
impact and are not limited to specific individuals or groups. This includes exploring large-scale
social and environmental challenges (Akula & Garibay, 2021; Holzmeyer, 2021; Tomašev et
al., 2020), such as how to manage natural (United Nations, 2023; Butler, 2017) and war
(Cornebise et al., 2018) disasters more effectively, minimize the effects of global warming, and
develop alternative energy sources (Chui et al., 2018).
Alongside this stream of research, there are AI studies that concentrate more on the well-
being of individuals and communities, particularly those who are vulnerable or marginalized.
Research on AI in social welfare is mostly dedicated to exploring possibilities for automation
and enhancement of social welfare systems (Oravec, 2019) comprised of diverse services,
policies, and programs aimed at ensuring that people's basic needs – such as health, education,
housing, employment, and income security – are met.
Another area of research that can be distinguished is focused on the application of AI in
providing specific social services in both the private and public sectors. This type of research
is concerned with how organizational structures implement specific interventions, programs,
practices, and work methods in assisting the functioning and well-being of individuals,
families, or communities. It is considered that AI could be applied in assessing eligibility and
needs, making enrolment decisions, providing benefits, and monitoring and managing the
delivery of benefits to clients of social services (Ohlenburg, 2020). AI is also being used to
conduct risk assessments, assist people in crisis, strengthen prevention efforts, identify
systemic biases in the delivery of social services, provide social work education, and predict
social worker burnout and service outcomes (Reamer, 2023). These benefits aligned with the
efforts of steady digital society development in the EU (Kersan-Škabić & Vukašina, 2023).
However, research on the application of AI in the field of social services is not widely
developed. In Lithuania, this is, on one hand, related to the lack of a strategically organized
plan for financing AI scientific research, and on the other hand, both in Lithuania and
worldwide, the development of AI systems in the social welfare does not proceed very rapidly.
AI development is privileged in other sectors that bring more financial benefit (Ministry of
Economy and Innovations, 2019). In this regard, proofs of positive impact of AI on business
development are obtained by Kolková & Ključnikov (2022); Letkovsky et al. (2023); Roshchyk
et al. (2022). Therefore, to promote future research on AI development and implementation in
the social service sector, critically evaluated and even competing conceptual frameworks are
needed, which will help in creating reliable knowledge.
1 „Artificial Intelligence (AI) refers to systems that display intelligent behavior by analyzing their environment and taking actions – with some degree of autonomy – to achieve specific goals. AI-based systems can be purely software-based, acting in the virtual world (e.g., voice
assistants, image analysis software, search engines, speech and face recognition systems), or AI can be embedded in hardware devices (e.g.,
advanced robots, autonomous cars, drones, or Internet of Things applications)“ (Ministry of Economy and Innovations, 2019). 2 The similar term 'AI for Good' refers to a United Nations project. It is the leading action-oriented, global, and inclusive platform on AI. Its
goal is to identify practical applications of AI that can advance the United Nations Sustainable Development Goals and scale those solutions
for global impact. (United Nations, 2022).
Seniutis, Gružauskas, Lileikiene, & Navickas
8
Current Research
Research on the development and application of AI for social welfare purposes has been
particularly expanding in recent years. For instance, computer science representatives like
Floridi et al. (2021) and Tomašev et al. (2020) discuss fundamental ethical principles and
practices in designing ethical AI or applying AI in the realm of social welfare. Application of
AI in law enforcement is analysed by Carton et al. (2016), Fang et al. (2016); in transportation
by Bojarski et al. (2016); in education by Lakkaraju et al. (2015); in knowledge management
by Bencsik (2021) and Bilan et al. (2023); in healthcare by Ross & Swetlitz (2017); Yu et al.
(2018) and Strickland (2019). While there are fewer studies focusing on the ethical aspects of
AI application in the social service sector or addressing specific social problems. In the field
of social sciences, management, and economics, Kim et al. (2022) provide guidelines for
creating AI implementation strategies in social innovation projects. Well-known cases involve
AI being used for predicting poverty risk (Jean et al., 2016), mandatory education for homeless
youth (Yadav et al., 2016), revealing factors contributing to criminal victimization among
homeless adults (Shah et al., 2021), automation of decisions made by child welfare specialists
(Gillingham, 2021).
In Lithuania, research is conducted on AI application in medical diagnosis (Janušonytė,
2021; Venclovaitė, D., Stramkauskaitė, A., Kuzmienė, 2022); the possible impact of AI on
different aspects of human consciousness is discussed from a philosophical perspective
(Vidauskytė, 2021); challenges raised by AI are analysed in constitutional law, the influence
of AI on human rights in law enforcement activities, etc. (Juškevičiūtė-Vilienė, 2020;
Skardžiūtė, 2018; Zakaras, 2022).
Previous research has primarily focused on general ethical aspects of AI rather than the
specific ethical challenges that may arise within the processes of social service delivery. This
involves the development of tailored AI solutions for socially vulnerable groups, ensuring their
unique needs are addressed while safeguarding their interests.
Research Objectives
The main aim of this study to devise a universally applicable conceptual framework for the
ethical implementation of artificial intelligence (AI) in social services, grounded in an
unstructured review of international scientific and normative literature on AI's application in
social welfare and its ethical governance. This framework, while globally relevant, will include
the backdrop of the Lithuanian context, offering insights into its practical application in a
specific national setting.
In order to achieve this aim, the study addresses the following objectives, which reflect the
main structure of this article: a) To analyse the growing expansion of AI as articulated in
political declarations and its implications for society; b) To investigate and document a range
of AI applications within the context of social welfare; c) To identify and scrutinize ethical
concerns associated with AI; d) To evaluate the current status of digitization in the Lithuanian
public sector; e) To delineate the requirements and directions for future research in the field of
ethical AI-based engineering of social services.
Conceptual Framework for Ethical Artificial Intelligence Development in Social Services Sector
9
Research Relevance It is expected that conceptual framework applied in future empirical research will provide key
insights and practical benefits for various stakeholders. Policymakers and service organizers
will gain guidance on key aspects like human resource investment, essential skills for AI
integration, impact measurement, risks management, and implementation of appropriate AI
applications. Service providers and practitioners will benefit from guidelines on maintaining
unbiased AI assessments, preserving professional autonomy, and maximizing the benefits of
AI. Additionally, service clients will be equipped with information and advice on handling
potential harm, ensuring transparent decision-making, and protecting their privacy in the
context of AI utilization.
METHODOLOGY
This study presents an unstructured internet sources literature review of 14 normative
documents on AI implementation, with a particular focus on international ethical regulations
regarding AI, and 40 scientific sources on Ethical AI applications for Social Good, with
particular attention to the field of social welfare (see references). This review can be
typologically aligned with an expository literature review, wherein social researchers engage
in scholarly discourse by offering a detailed explanation of the subject matter. They utilize pilot
evidence to foster a comprehensive understanding of the topic under examination (Tayo et al.,
2023). This research method was chosen because it grants access to existing knowledge,
offering solutions or suggestions aligned with the main research aim, and it also guides the
research towards achieving its findings (Yekeen, 2006). The main stages of review such as
formulation of selection criteria, selection of sources, data extraction and analysis were
performed according to guidelines of Kitchenham and Charters (2007).
The selection criteria for normative literature sources were established to identify
documents that regulate AI implementation across multiple levels: nationally in Lithuania,
internationally within the European Union and the United States, and globally under United
Nations guidelines. Additionally, for scientific sources, the focus was on selecting articles that
explore AI implementation in areas such as social good initiatives, social welfare, and public
The selection criteria for normative literature sources services, with particular attention to the
ethical challenges presented in these areas. As for each qualitative study, it was important that
selected sources would be representative in relation to the research problem and not to the
volume of available literature or the prevalence of certain viewpoints (Židžiūnaitė &
Sabaliauskas, 2017).
To search for relevant normative literature, we utilized Google Search Engine. For
scientific literature, our approach included accessing databases with open access, such as
Google Scholar, Scopus, and Web of Science using different combination of these keywords:
AI, Ethics, Regulation, Social Good, Social Welfare, Social and/or Public Service.
Data analysis in our study was performed through qualitative thematic analysis using
MAXQDA 2022 software. This approach enabled us to identify and organize key themes that
are crucial to achieving the study’s objectives systematically and inductively. Data analysis
was conducted according to the following procedure: familiarization with research data, data
Seniutis, Gružauskas, Lileikiene, & Navickas
10
coding, searching for themes, reviewing them, description, and report preparation (Broun &
Clarke, 2014). These main themes are integral not only to the study's structure, reflected in its
objectives and section titles, but they also form the core categories of our conceptual
framework. These themes include: 1) Applications of AI; 2) AI's impact on various social
service processes; 3) the process of delivering social services; 4) Ethical issues in AI usage;
and 5) Strategies for risk prevention and resolution in the context of AI implementation in
social services.
ARTIFICIAL INTELLIGENCE IN SOCIAL WELFARE: A LITERATURE REVIEW
Rising Expansion and Expectations Toward AI for Society in Political Declarations
There is a growing number of political declarations that not only encourage the development
of artificial intelligence (AI) globally but also express the expectation that AI will contribute
to social good in a broad sense and ensure social welfare specifically.
Documents from the European Commission and various international bodies over the last five
years highlight substantial aspirations regarding the advancement of AI. Firstly, there is a
strong ambition for the European Union (EU) to become a world leader in both the
development and application of AI. It is expected that each member of the EU will contribute
towards establishing the EU as a champion of an AI approach that benefits both individuals
and society (European Commission, 2018), there is an optimistic view that AI will significantly
contribute to creating a more sustainable society and drive economic growth. This includes
improving the efficiency of healthcare and agriculture, aiding in the reduction of climate
change, enhancing the productivity of manufacturing systems, and strengthening security
measures (European Commission, 2020). The goal set for 2030 is for 75% of European
companies to integrate cloud computing, big data, and artificial intelligence technologies. The
expectation is that the adoption of AI will lead to improvements in job quality, workplace
safety, efficiency, and employee well-being (European Commission, 2021a). Thirdly, one of
the EU digital policy objectives for the coming decade is the development of Ethical Artificial
Intelligence. This aims to foster responsible and reliable AI that benefits humanity and upholds
human rights globally (European Commission, 2021c). Similarly, the Global Goals established
by the United Nations in 2015 emphasize that AI should play a crucial role in advancing social
welfare, protecting human rights, and promoting environmental sustainability, among other
objectives (United Nations, 2015).
Thus, the ambition to become a leader in the AI industry, coupled with elevated
expectations for AI's positive impact across various social and economic areas of life, and the
simultaneous focus on the ethical challenges of AI implementation, may be considered key
elements of the EU's political agenda related to AI development.
Exploring the Range of AI Applications in Social Welfare
Scientific studies confirm that expectations towards AI are being, to some extent, realized.
Various AI applications have demonstrated success in solving a range of social problems and
Conceptual Framework for Ethical Artificial Intelligence Development in Social Services Sector
11
in delivering diverse social services within specific institutional frameworks, thereby
contributing to the enhancement of the social welfare system (Shi et al., 2020).
Illustrative of this are the following applications: a) Natural Language Processing (NLP)
– algorithms capable of reading and comprehending human language have been adapted as
virtual assistants for social service users, practitioners, or administrators (Chui et al., 2023); b)
Computer Vision (CV) – algorithms that process visual information can enhance public safety
by detecting instances of violence, locating missing persons (Trilupaitytė, 2022), aiding the
visually impaired (Patel & Parmar, 2022), and assisting seniors with dementia (Bucholc et al.,
2023; Chignell et al., 2020); c) Robotics – mobile systems capable of autonomous movement
and environmental interaction (Burgard, 2023) can significantly aid in the care and support of
the elderly (Wellman & Rajan, 2017); d) Machine Learning (ML) – this method uses historical
or real-time data to predict future trends (Engin & Treleaven, 2019), assisting in identifying
vulnerable tenant issues (Yeung, 2018), modelling migration patterns (Robinson & Dilkina,
2018), supporting the homeless (Abelson et al., 2014), developing victimization models (Shah
et al., 2021), and optimizing food distribution and usage (Shi et al., 2020).
The presented classification of various AI applications, or in other words, the breadth of
AI industries, is neither definitive nor exhaustive regarding the fields of its application.
However, this overview offers essential insights into the already achieved development of AI
in the field of social services in different countries around the world.
Identifying Ethical AI issues However, numerous ethical challenges arise in relation to the implementation of mentioned AI
applications in various socio-economic contexts (United Nations, 2021). These potential
ethical challenges are well-summarized in the White House's (2022) Blueprint for AI-related
legislation and ethical considerations.
The ethical challenges linked to AI application encompass a spectrum of issues. Firstly,
there are security risks and effectiveness concerns, where AI applications might endanger users'
safety and amplify deception, affecting trust and reliability. Secondly, algorithmic bias presents
a problem, as automated algorithms could inadvertently favor certain demographics, leading to
inequality based on age, gender, ethnicity, and more. Thirdly, data privacy issues arise from
excessive data collection and use without explicit consent or for unspecified purposes.
Additionally, the opacity of AI systems makes their results and processes often difficult to
comprehend and verify. Lastly, there's the issue of autonomy loss, where reliance on AI leaves
no alternative options for people, limiting their freedom of choice.
The ethical guidelines promoted by the (European Commission, 2019) highlight seven key
requirements for AI systems to be deemed trustworthy. Additionally, the document “Artificial
Intelligence Act” (European Commission, 2021b) presents a more developed and
comprehensive treatment of AI requirements. Many of these requirements are in line with the
previously presented White House AI regulations, for example: a) Technical robustness and
safety, including resilience to attack and security, a fall-back plan and general safety, accuracy,
reliability, and reproducibility. b) Diversity, non-discrimination, and fairness, encompassing
the avoidance of unfair bias, accessibility and universal design, and stakeholder participation.
c) Privacy and data governance, which covers respect for privacy, quality and integrity of data,
and access to data. d) Transparency, including traceability, explainability, and communication.
Seniutis, Gružauskas, Lileikiene, & Navickas
12
e) Human agency and oversight, ensuring fundamental rights, human agency, and human
oversight.
The additional requirements that are less developed in United States regulation include
Societal and Environmental Well-being, which mandates that the entire supply chain of AI
systems must be environmentally friendly and beneficial to all human beings, including future
generations. It involves a critical examination of resource usage and energy consumption,
particularly during the AI systems' training phase. Additionally, AI systems must ensure a
positive impact on the social relationships, physical, and mental well-being of users. It is also
essential to assess their influence on political processes and structures, including political
decision-making and electoral contexts.
One more less developed requirement is accountability, which primarily entails the
establishment of mechanisms that enable the auditing of AI systems. This means evaluating
and assessing the processes and outcomes of AI systems' algorithms. AI systems should be
subject to independent and continuous auditing throughout their entire lifecycle to ensure their
safety. AI developers must implement risk and impact assessment procedures, and AI users
should have the opportunity to report any negative impacts they might encounter.
The competition for leadership in the AI industry inevitably leads to joint efforts to define
ethical principles for AI application and to formulate corresponding AI requirements. Both the
European Commission and the White House have promoted a human-centric approach to AI,
aiming to enhance individual and collective human well-being. This leads to the fostering of
elaboration and international consensus on AI ethics.
Status Quo of the Digitization of the Lithuanian Public Sector
The digitalization of public services (health, education, legal services, etc.) in Lithuania is
progressing rapidly. According to the Digital Economy and Society Index (DESI) (European
Commission, 2022) Lithuania ranks tenth – 83,85 (score 0 to 10) among European Union
countries in this regard. The EU average is – 77, 03 (score 0 to 100) (see figure 1).
However, the pace of AI development in the public sector is not particularly rapid, which
could be attributed to the others strategic planning priorities and to the relatively modest
financial resources dedicated to AI development in Lithuania. From 2015 to 2018, the public
sector invested 26.5 million euros in AI development. For comparison, France's AI
development plan is valued at 1.5 billion euros, and China's plan is worth 150 billion euros
(Ministry of Economy and Innovations, 2019).
There are known cases where AI systems are applied in law enforcement institutions:
police, border guard services, prison department, and other state institutions (Zakaras, 2022),
as well as in the health sector, for example in diagnosing diseases (Janušonytė, 2021;
Venclovaitė, D., Stramkauskaitė, A., Kuzmienė, 2022) and the business sector (European
Commission, 2022), but there are no cases of AI application in the social services sector. It
seems that AI development in Lithuania is prioritized in the financial sector. Only 4.45 % of
all enterprises (excluding the financial sector) have applied artificial intelligence, whereas the
European Union average is 7.91 % (see figure 2).
Conceptual Framework for Ethical Artificial Intelligence Development in Social Services Sector
13
Figure 1. Digital public service for citizens (score 0 to 100), DESI 2022 (data from 2021)
Figure 2. AI implementation across all enterprises (employing 10 or more people and excluding the
financial sector) DESI 2022 (data from 2021)
In the Lithuanian Artificial Intelligence Strategy, initiated by the Ministry of Economy and
Innovation in 2018-2019, multiple sectors (see figure 3) were pinpointed for prioritized AI
development, chosen based on their economic benefit and sectoral impact. The strategy targets
a) Manufacturing, aiming to boost productivity through automation; b) Agriculture, employing
AI for tasks like robotic harvest collection and intelligent soil analysis; c) Transport, leveraging
AI for traffic management and autonomous vehicles; d) Energy, utilizing AI for more efficient
Seniutis, Gružauskas, Lileikiene, & Navickas
14
energy supply methods, reducing foreign energy dependence; and e) Healthcare, integrating AI
to manage increasing patient loads and paperwork, thereby optimizing healthcare delivery.
Figure 3. Priority sectors for AI development in Lithuania
Source: own elaboration.
Except the health sector, others are not directly related to social welfare. Therefore, it can
be stated that the goals of social welfare are not sufficiently expressed in Lithuania's AI
strategy. In a broader sense, the implementation of AI in all these sectors could eventually in
long term contribute to a better quality of life of entire population. However, the urgent needs
of individuals and groups in need, for whom the social welfare system is designed to provide
assistance in various forms, including financial support, healthcare, education, housing, and
other services, seem to be insufficiently relevant and prospective to be considered in planning
AI development in the country.
In the main strategic documents concerning the state's social welfare, such as the
Lithuanian Artificial Intelligence Strategy by the Ministry of Economy and Innovations (2019)
and the Strategic Plan by the Ministry of Social Security and Labour (2023), there is a lack of
mention of AI implementation in the field of social services. Overall, such plans will have to
emerge, along with legal regulation, funding, and specific attempts to create and apply AI in
the field of social services. It remains to wish that this process at the political decision-making
level would go as smoothly as possible, and at the practical level, opportunities would be
created to prepare for the upcoming innovations, such as professional training programs on
artificial intelligence, publicly accessible resources about AI opportunities and risks, scientific
research, and the like.
Conceptual Framework for Ethical Artificial Intelligence Development in Social Services Sector
15
RESULTS
Conceptual Framework for Integrating Ethical AI into the Social Service Process There are well-known national initiatives that involve collaborations between the government,
industry (AI companies), researchers (including computer engineers and social scientists), and
practitioners (such as social service organizers and providers) (Fox et al., 2022; Bartosz
Gajderowicz, 2019). These initiatives are aimed at developing social service solutions based
on AI systems. Interdisciplinary and cross-sectoral expert groups are working on social service
engineering, which includes how each stage of the social services chain can benefit from
engineering design, planning, and delivery. The inclusion of AI systems in this engineering
process has the potential to enhance the effectiveness and efficiency of social services. This is
achieved by delivering the right services to the right people at the right time, and by preventing
and/or addressing potential ethical challenges. For example, a total of $4.9 million in funding
was allocated for the implementation of the Compass project in Canada. The goal is to develop
a national AI-based platform that aligns the needs of individuals, families, and communities
with the right combination of social service options (Darling, 2022).
In this study, a proposed conceptual framework integrates, details, and illustrates potential
relationships among four main categories: social service process, the impact of AI on various
stages in social service process, AI applications, ethical issues, and risk prevention and solution
(see figure 4).
Firstly, the social service process is viewed as a value chain, identifying universal stages
such as: design of social services, access to services; identification of client needs; planning
assistance; delivery of social services; evaluating service effectiveness (see Saunders, 2016).
Secondly, AI systems can be integrated into various stages of the social service process,
potentially having a significant impact on these stages. a) AI can analyse large-scale data to
simulate clients’ interactions in social services identifying trends and gaps. This helps the
design of services focused on real clients’ needs and proven practices (Gajderowicz et al.,
2014). b) It can provide personalized guidance for users, helping them navigate and access
appropriate services more efficiently (Vendeville, 2022). c) AI's data analysis capabilities
enhance the precise identification of individual needs by emulating and analysing client
behaviour. This understanding of specific client behaviours then guides the customization of
suitable assistance methods (Gajderowicz & Fox, 2017). d) AI may enhance the process of
searching for and aligning appropriate services with client needs, as well as predicting the
potential success rates of these services (Rosu et al., 2017). e) AI can automate routine tasks
and decision-making process in service delivery (Jankovic & Bernier, 2023). f) AI can enhance
the evaluation process through advanced data analytics, thereby assisting managers and policy
makers in improving social services and policies (Vendeville, 2022).
Thirdly, a variety of AI applications, including natural language processing, computer
vision, robotics, and machine learning, are examined for their alignment with each stage of the
service process (see section 2).
Subsequently, ethical challenges are identified, encompassing issues such as security risks
and effectiveness, algorithmic bias, data privacy issues, opacity, autonomy loss, societal and
environmental well-being, accountability (see section 3).
Seniutis, Gružauskas, Lileikiene, & Navickas
16
Figure 4. Conceptual Framework for Ethical AI-based social service engineering
Source: own elaboration.
Conceptual Framework for Ethical Artificial Intelligence Development in Social Services Sector
17
Finally, opportunities for preventing or addressing these challenges are highlighted,
including strategies like a) cyber security – the practice of protecting computer systems,
networks, and data from digital attacks, theft, or damage. It encompasses technologies,
processes, and controls designed to defend against cyber threats (Hendrycks et al., 2022; Zhou
et al., 2023); b) data quality and diversity – refers to the accuracy, completeness, reliability,
and relevance of data in AI systems. Diversity in data ensures that AI algorithms are trained on
a wide range of datasets to avoid bias and to perform effectively across different scenarios
(Suresh & Guttag, 2021; Torralba & Efros, 2011); c) identification & depersonalization –
identification in AI involves recognizing and distinguishing individual entities, whereas
depersonalization is the process of removing personally identifiable information from data sets,
ensuring privacy and anonymity (Lison et al., 2021; Patsakis & Lykousas, 2023); d)
explainable artificial intelligence (XAI) – the field of AI that focuses on the creation of AI
systems whose actions can be easily understood by humans. XAI aims to make AI decisions
transparent, understandable, and interpretable (Barredo Arrieta et al., 2020; Molnar, 2019); e)
human involvement (human-in-the-loop) – a system design paradigm that incorporates human
judgment into AI systems, allowing humans to provide feedback, make decisions, or adjust
outputs in real-time, ensuring that the AI remains aligned with human values and goals
(Mullainathan & Obermeyer, 2017; Shevlane et al., 2023); f) alternative energy sources /
quantum computing – this refers to the exploration and use of renewable energy sources, like
solar or wind power, to run AI computations, and the application of quantum computing to
dramatically increase computational power for certain types of problems, potentially
improving AI efficiency and capabilities (Ajagekar & You, 2019; Jaschke & Montangero,
2023; Li et al., 2022; McDonald et al., 2022); g) content moderation based on free choice – n
AI allows users to customize content filters to their ethical preferences and cultural norms,
offering a personalized approach to blocking or allowing content instead of a universal
moderation policy. This model promotes a balance between preventing harm and upholding
diverse expressions (Bubeck et al., 2023; Open AI, 2023); h) monitoring – the continuous
observation, checking, and tracking of AI systems' performance and activities. Monitoring aims
to ensure that AI systems function as intended, remain secure, and any issues are promptly
identified and addressed (Sculley et al., 2015).
This Conceptual Framework presents a general model for evaluating the impact of diverse
AI applications across the social service process, pinpointing key ethical concerns that may
emerge. It underscores the necessity of incorporating ethical considerations at the initial stages
of AI development. While the framework is universally applicable, it is especially pertinent in
examining the prospective integration of AI within Lithuania's social service sector. This case
study approach offers valuable perspectives on the effective and ethical implementation of AI.
The framework's alignment with standard social service stages ensures its adaptability for
investigating the potential deployment of various AI systems in delivering a range of social
services to different client groups in varied institutional settings.
Seniutis, Gružauskas, Lileikiene, & Navickas
18
CONCLUSION
Human-centered ethical AI development, guided by international efforts, is a priority yet
remains an unachieved goal. The leadership of the EU and USA in AI regulation and the AI
industry raises expectations for AI's positive social and economic impact. Research shows that
AI's technological breakthrough has successfully applied many AI systems in the field of social
welfare. However, in Lithuania, strategic social welfare documents lack specific AI integration
plans, highlighting the need for legal regulation, funding, and AI strategy development in social
services. The Conceptual Framework proposed in this research provides critical insights for the
ethical and effective integration of AI into social service sectors, exemplified through the
Lithuanian context but designed for universal applicability. Moreover, the universality of social
service stages enhances the framework's utility in applying it to diverse AI applications across
various client groups and institutional settings.
The proposed framework delineates a methodical integration of AI into social service
delivery, highlighting the transformative potential of technologies such as natural language
processing, computer vision, robotics, and machine learning. These technologies, when
judiciously applied, can streamline the processing of information and enhance client
interactions through automation, offering personalized and efficient service provision. Delving
into the social service stages, the framework elucidates how AI can critically reformulate the
lifecycle of services—from the design of predictive and adaptive social programs, to improving
accessibility with intelligent guidance systems, from harnessing big data for need identification
and planning, to the alignment of services with client requirements through smart matching
algorithms. At the ethical forefront, the framework insists on the rectification of algorithmic
biases by curating diverse datasets for training AI systems. It underlines the imperative of
safeguarding data privacy, asserting transparency in the workings of AI, and upholding
accountability for AI-induced decisions, with an insistent recommendation for human
oversight in key decision-making junctures. Risk prevention is addressed through a set of
practical guidelines that encompass stringent cybersecurity protocols, the enhancement of data
quality, the promotion of AI explainability, and the integration of humans in the loop,
particularly in sensitive decision-making processes. For pragmatic enactment, the framework
advocates regular auditing, stakeholder education, and community engagement, reinforcing the
role of AI in social services as a collaborative nexus of technology, ethics, and human welfare.
In the conclusive synthesis of our research, it is imperative to translate the theoretical
underpinnings of our framework into actionable recommendations. These practical suggestions
are intended to guide key stakeholders—policymakers, service providers, clients, and the
community at large—on the responsible integration of AI into the social services sector. For
policymakers, the framework recommends the development of robust AI governance models
that not only foster innovation but also prioritize ethical considerations. This includes creating
policies that encourage the ethical collection and use of data, implementing standards for
transparency, and promoting equitable access to AI-enhanced services. Service providers are
encouraged to adopt AI solutions that are congruent with the core values of social work,
ensuring that such technologies are used to augment, rather than replace, human judgment and
empathy. Service providers should also focus on training their staff to work effectively with
AI tools and to understand their capabilities and limitations. Clients, as end-users of social
services, should be educated on how AI may impact their access to and the quality of services
Conceptual Framework for Ethical Artificial Intelligence Development in Social Services Sector
19
provided. They should be empowered with the knowledge to navigate AI-driven services safely
and with awareness of their rights, particularly regarding data privacy. For other involved
parties, including technologists, ethicists, and social welfare experts, the framework advises a
collaborative approach to ensure that AI technologies are developed and implemented in a
socially responsible manner. This includes continuous dialogue on the ethical implications of
AI, shared responsibility for risk mitigation, and a commitment to the ongoing evaluation of
AI's impact on social welfare.
REFERENCES
Normative documents
European Commission. (2018, April 20). Artificial Intelligence for Europe. https://eur-lex.europa.eu/legal-
conhttps://eur-lex.europa.eu/legal-content/EN/TXT/HTML/?uri=CELEX:52018DC0237
European Commission. (2019, April 8). Ethics Guidelines for Trustworthy AI. High-Level Expert Group on
Artificial Intelligence. https://ec.europa.eu/futurium/en/ai-alliance-consultation.1.html#:~:text=The AI
HLEG presented, and published in April 2019
European Commission. (2020, February 19). White Paper on Artificial Intelligence a European approach to
excellence and trust. https://ec.europa.eu/commission/sites/beta-political/files/political-guidelines-next-
commission_en.pdf.
European Commission. (2021a, March 9). 2030 Digital Compass: the European way for the Digital Decade.
https://eur-lex.europa.eu/legal-content/en/TXT/?uri=CELEX%3A52021DC0118
European Commission. (2021b, April 21). Artificial Intelligence Act. https://eur-lex.europa.eu/legal-
content/EN/TXT/?uri=celex%3A52021PC0206
European Commission. (2021c, April 21). Fostering a European approach to Artificial Intelligence. https://eur-
lex.europa.eu/legal-content/EN/ALL/?uri=COM%3A2021%3A205%3AFIN
European Commission. (2022). The Digital Economy and Society Index. https://digital-
strategy.ec.europa.eu/en/policies/desi
United Nations. (2015). Transforming our world: the 2030 Agenda for Sustainable Development.
https://www.refworld.org/docid/57b6e3e44.html
United Nations. (2021). Recommendation on the Ethics of Artificial Intelligence (Issue November).
https://doi.org/10.7551/mitpress/14102.003.0010
United Nations. (2022). United Nations Activities on Artificial Intelligence (AI). https://eur-lex.europa.eu/legal-
content/EN/TXT/HTML/?uri=CELEX:52018DC0237
United Nations. (2023). Building fertile ground for data science in Uganda. United Nations Global Pulse.
https://medium.com/un-global-pulse/building-fertile-ground-for-data-science-in-uganda-a950dfd3ca0b
White House. (2022). Blueprint for an AI Bill of Rights. Making automated systems work for the American
peopleTEMS WORK FOR THE AMERICAN PEOPLE. https://www.whitehouse.gov/ostp/ai-bill-of-rights
Ministry of Economy and Innovations. (2019). Lithuanian artificial intelligence strategy: a vision for the future.
Ministry of Social Security and Labour. (2023). Strategic Plan.
Seniutis, Gružauskas, Lileikiene, & Navickas
20
Scientific sources
Abelson, B., Varshney, K. R., & Sun, J. (2014). Targeting direct cash transfers to the extremely poor. Proceedings
of the ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, 1563–1572.
https://doi.org/10.1145/2623330.2623335
Akula, R., & Garibay, I. (2021). Ethical AI for Social Good. Lecture Notes in Computer Science (Including
Subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), 13095 LNCS, 369–
380. https://doi.org/10.1007/978-3-030-90963-5_28/COVER
Bucholc, M., James, C., Al Khleifat, A., Badhwar, A., Clarke, N., Dehsarvi, A., Madan, C. R., Marzi, S. J., Shand,
C., Schilder, B. M., Tamburin, S., Tantiangco, H. M., Llewellyn, D. J., & Ranson, J. M. (2023). Artificial
intelligence for dementia research methods optimization The Deep Dementia Phenotyping (DEMON)
Network 1 Ilianna Lourida 14. https://doi.org/10.1002/alz.13441
Burgard, W. (2023). Artificial Intelligence: Key Technologies and Opportunities. In K. Frankish & W. M. Ramsey
(Eds.), The Cambridge handbook of artificial intelligence (pp. 11–19). Cambridge University Press.
Chignell, M., Matulis, H., & Nejati, B. (2020). Motivating Physical Exercise in the Elderly with Mixed Reality
Experiences. In N. Streitz & S. Konomi (Eds.), 8th International Conference Distributed, Ambient and
Pervasive Interactions (DAPI) 2020: Vol. 12203 LNCS (pp. 505–519). Springer.
https://doi.org/10.1007/978-3-030-50344-4_36/TABLES/3
Chui, M., Harryson, M., Manyika, J., Roberts, R., Chung, R., van Heteren, A., & Nel, P. (2018). Notes from the
AI frontier. Applying AI for Social Good.
Chui, M., Issler, M., Roberts, R., & Yee, L. (2023). McKinsey Technology Trends Outlook 2023.
https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/the-top-trends-in-tech#tech-trends-
2023
Cornebise, J., Worrall, D., Farfour, M., & Marin, M. (2018). Witnessing atrocities: quantifying villages
destruction in Darfur with crowdsourcing and transfer learning. AI for Social Good NeurIPS2018 Workshop,
NeurIPS ’18,
Darling, E. (2022). Digital Technology Supercluster Announces Investment to Improve the Accessibility of Social
Services. https://www.digitalsupercluster.ca/digital-technology-supercluster-announces-investment-to-
improve-the-accessibility-of-social-services/
Engin, Z., & Treleaven, P. (2019). Algorithmic Government: Automating Public Services and Supporting Civil
Servants in using Data Science Technologies. Advance Access Publication On, 11.
https://doi.org/10.1093/comjnl/bxy082
Floridi, L., Cowls, J., King, T. C., & Taddeo, M. (2021). How to Design AI for Social Good: Seven Essential
Factors. Philosophical Studies Series, 144, 125–151. https://doi.org/10.1007/978-3-030-81907-1_9/COVER
Fox, M. S., Gajderowicz, B., Rosu, D., Turner, A., & Lyu, D. (2022). An Ontological Approach to Analysing
Social Service Provisioning. ISC2 2022 - 8th IEEE International Smart Cities Conference.
https://doi.org/10.1109/ISC255366.2022.9922132
Gajderowicz, Bart, & Fox, M. S. (2017). Requirements for Emulating Homeless Client Behaviour. Conference:
AAAI Workshop on Operations Research and Artificial Intelligence for Social Good.
https://www.researchgate.net/publication/313556478
Gajderowicz, Bart, Fox, M. S., & Grüninger, M. (2014). Requirements for an Ontological Foundation for
Modelling Social Service Chains. Proceedings of the 2014 Industrial and Systems Engineering Research
Conference Y. Guan and H. Liao, Eds.
Gajderowicz, Bartosz. (2019). Artificial Intelligence Planning Techniques for Emulating Agents with Application
in Social Services. University of Toronto (Canada).
Gillingham, P. (2021). Algorithmically Based Decision Support Tools: Skeptical Thinking about the Inclusion of
Previous Involvement. Practice, 33(1), 37–50. https://doi.org/10.1080/09503153.2020.1749584
Conceptual Framework for Ethical Artificial Intelligence Development in Social Services Sector
21
Holzmeyer, C. (2021). Beyond ‘AI for Social Good’ (AI4SG): social transformations—not tech-fixes—for health
equity. Interdisciplinary Science Reviews, 46(1–2), 94–125.
https://doi.org/10.1080/03080188.2020.1840221
Janušonytė, E. (2021). Odos vėžio diagnostika ir dirbtinis intelektas. Sveikatos Mokslai, 31(3), 175–180.
https://doi.org/10.35988/sm-hs.2021.103
Jean, N., Burke, M., Xie, M., Davis, W. M., Lobell, D. B., & Ermon, S. (2016). Combining satellite imagery and
machine learning to predict poverty. Science, 353(6301), 790–794.
https://doi.org/10.1126/SCIENCE.AAF7894/SUPPL_FILE/JEAN.SM.PDF
Juškevičiūtė-Vilienė, A. (2020). Dirbtinis intelektas ir konstitucinė teisė į teisingumą. Acta Universitatis
Lodziensis. Folia Iuridica, 93, 117–136.
Kim, E., Jang, G. Y., & Kim, S. H. (2022). How to Apply Artificial Intelligence for Social Innovations. Applied
Artificial Intelligence, 36(1). https://doi.org/10.1080/08839514.2022.2031819
Oravec, J. A. (2019). Artificial Intelligence, Automation, and Social Welfare: Some Ethical and Historical
Perspectives on Technological Overstatement and Hyperbole. Ethics and Social Welfare, 13(1), 18–32.
https://doi.org/10.1080/17496535.2018.1512142
Patel, K., & Parmar, B. (2022). Assistive device using computer vision and image processing for visually
impaired; review and current status. Disability and Rehabilitation: Assistive Technology, 17(3), 290–297.
https://doi.org/10.1080/17483107.2020.1786731
Reamer, F. G. (2023). Artificial Intelligence in Social Work: Emerging Ethical Issues. International Journal of
Social Work Values and Ethics, 20(2), 52–71. https://doi.org/10.55521/10-020-205
Robinson, C., & Dilkina, B. (2018). A Machine Learning Approach to Modeling Human Migration. COMPASS
’18: Proceedings of the 1st ACM SIGCAS Conference on Computing and Sustainable Societies, 1–9.
https://doi.org/10.1145/3209811.3209868
Rosu, D., Aleman, D. M., Beck, J. C., Chignell, M., Consens, M., Fox, M. S., Grüninger, M., Grüninger, G., Liu,
C., Ru, Y., & Sanner, S. (2017). A Virtual Marketplace for Goods and Services for People with Social Needs.
Saunders, R. P. (2016). Implementation Monitoring and Process Evaluation.
Shah, O. R., Willoughby, L., & Bowersox, N. (2021). Issues in Information Systems Tackling homelessness
through AI powered social innovations: A novel and ground-breaking assessment of criminal victimization
of homeless populations in los angeles employing predictive analytics and machine learning models such as
ARIMA and LSTM. 22(3), 264–277. https://doi.org/10.48009/3_iis_2021_283-297
Shi, Z. R., Wang, C., & Fang, F. (2020). Artiicial Intelligence for Social Good: A Survey. Preprint:
ArXiv:2001.01818. https://aaai.org/Symposia/Spring/sss17symposia.php
Skardžiūtė, J. (2018). Dirbtinio intelekto įtaka konkurencijos teisei: konkurencijos teisės normų aiškinimo ir
taikymo problemos [Vilniaus universitetas]. https://epublications.vu.lt/object/elaba:29809699/
Tomašev, N., Cornebise, J., Hutter, F., Mohamed, S., Picciariello, A., Connelly, B., Belgrave, D. C. M., Ezer, D.,
Haert, F. C. van der, Mugisha, F., Abila, G., Arai, H., Almiraat, H., Proskurnia, J., Snyder, K., Otake-
Matsuura, M., Othman, M., Glasmachers, T., Wever, W. de, … Clopath, C. (2020). AI for social good:
unlocking the opportunity for positive impact. Nature Communications 2020 11:1, 11(1), 1–6.
https://doi.org/10.1038/s41467-020-15871-z
Trilupaitytė, S. (2022). Vizualioji kontrolė šiandienos visuomenėse: veidų ir emocijų (ne)atpažinimas.
Politologija, 2(106), 131–164.
Venclovaitė, D., Stramkauskaitė, A., Kuzmienė, L. (2022). Dirbtinis intelektas oftalmologijoje. Artificial
intelligence in ophthalmology. https://doi.org/10.37499/LBPG.942
Vendeville, G. (2022). Using AI to optimize social services: Professor Mark Fox among U of T researchers to
team up with industry and government. https://www.mie.utoronto.ca/using-ai-to-optimize-social-services-
professor-mark-fox-among-u-of-t-researchers-to-team-up-with-industry-and-government/
Seniutis, Gružauskas, Lileikiene, & Navickas
22
Vidauskytė, L. (2021). Dirbtinis intelektas: visuomenės infantilizacija ir bejėgiškumas. Logos (Vilnius), 109, 71–
77. https://doi.org/10.24101/LOGOS.2021.77
Wellman, M. P., & Rajan, U. (2017). Ethical Issues for Autonomous Trading Agents. Minds and Machines, 27.
https://doi.org/10.1007/s11023-017-9419-4
Yadav, A., Chan, H., Jiang, A., Rice, E., Kamar, E., Grosz, B., & Tambe, M. (2016). POMDPs for assisting
homeless shelters – Computational and deployment challenges. International Conference on Autonomous
Agents and Multiagent Systems, 10003 LNAI, 67–87. https://doi.org/10.1007/978-3-319-46840-
2_5/FIGURES/14
Yeung, K. (2018). Algorithmic regulation: A critical interrogation. Regulation & Governance, 10, 505–523.
https://doi.org/10.1111/rego.12158
Yu, K. H., Beam, A. L., & Kohane, I. S. (2018). Artificial intelligence in healthcare. Nature Biomedical
Engineering 2018 2:10, 2(10), 719–731. https://doi.org/10.1038/s41551-018-0305-z
Zakaras, K. (2022). Dirbtinio intelekto naudojimo teisėsaugos institucijų veikloje iššūkiai žmogaus teisių
požiūriu. Mykolo Romerio universitetas.
Secondary sources
Ajagekar, A., & You, F. (2019). Quantum computing for energy systems optimization: Challenges and
opportunities. Energy, 179, 76–89. https://doi.org/10.1016/j.energy.2019.04.186
Barredo Arrieta, A., Díaz-Rodríguez, N., Del Ser, J., Bennetot, A., Tabik, S., Barbado, A., Garcia, S., Gil-Lopez,
S., Molina, D., Benjamins, R., Chatila, R., & Herrera, F. (2020). Explainable Artificial Intelligence (XAI):
Concepts, taxonomies, opportunities and challenges toward responsible AI. Information Fusion, 58, 82–115.
https://doi.org/10.1016/J.INFFUS.2019.12.012
Bencsik, A. (2021). The sixth generation of knowledge management – the headway of artificial intelligence.
Journal of International Studies, 14(2), 84-101. doi:10.14254/2071-8330.2021/14-2/6
Bilan, Y., Oliinyk, O., Mishchuk, H., & Skare, M. (2023). Impact of information and communications technology
on the development and use of knowledge. Technological Forecasting and Social Change, 191, 122519.
DOI: 10.1016/j.techfore.2023.122519
Bojarski, M., Testa, D. Del, Dworakowski, D., Firner, B., Flepp, B., Goyal, P., Jackel, L. D., Monfort, M., Muller,
U., Zhang, J., Zhang, X., Zhao, J., & Zieba, K. (2016). End to End Learning for Self-Driving Cars.
Bubeck, S., Chandrasekaran, V., Eldan, R., Gehrke, J., Horvitz, E., Kamar, E., Lee, P., Lee, Y. T., Li, Y.,
Lundberg, S., Nori, H., Palangi, H., Ribeiro, M. T., & Zhang, Y. (2023). Sparks of Artificial General
Intelligence: Early experiments with GPT-4. Microsoft Research. https://arxiv.org/abs/2303.12712v5
Carton, S., Helsby, J., Joseph, K., Mahmud, A., Park, Y., Walsh, J., Cody, C., Patterson, C. P. T. E., Haynes, L.,
& Ghani, R. (2016). Identifying police officers at risk of adverse events. Proceedings of the ACM SIGKDD
International Conference on Knowledge Discovery and Data Mining, 13-17-August-2016, 67–76.
https://doi.org/10.1145/2939672.2939698
Fang, F., Nguyen, T. H., Pickles, R., Lam, W. Y., Clements, G. R., An, B., Singh, A., Tambe, M., & Lemieux, A.
(2016). Deploying PAWS: Field Optimization of the Protection Assistant for Wildlife Security. Proceedings
of the AAAI Conference on Artificial Intelligence, 30(2), 3966–3973.
https://doi.org/10.1609/AAAI.V30I2.19070
Hendrycks, D., Google, N. C., Schulman, J., & Steinhardt, J. (2022). Unsolved Problems in ML Safety. Preprint.
https://arxiv.org/abs/2109.13916
Jankovic, J., & Bernier, A. (2023). Research shows decision-making AI could be made more accurate when
judging humans. https://www.utoronto.ca/news/research-shows-decision-making-ai-could-be-made-more-
accurate-when-judging-humans
Conceptual Framework for Ethical Artificial Intelligence Development in Social Services Sector
23
Jaschke, D., & Montangero, S. (2023). Is quantum computing green? An estimate for an energy-efficiency
quantum advantage. Quantum Sci. Technol, 8, 25001. https://doi.org/10.1088/2058-9565/acae3e
Kersan-Škabić, I., & Vukašina, M. (2023). Contribution of ESIFs to the digital society development in the EU.
Journal of International Studies, 16(2), 195-210. doi:10.14254/2071-8330.2023/16-2/13
Kolková, A., & Ključnikov, A. (2022). Demand forecasting: AI-based, statistical and hybrid models vs practice-
based models - the case of SMEs and large enterprises. Economics and Sociology, 15(4), 39-62.
doi:10.14254/2071- 789X.2022/15-4/2
Lakkaraju, H., Aguiar, E., Shan, C., Miller, D., Bhanpuri, N., Ghani, R., & Addison, K. L. (2015). A machine
learning framework to identify students at risk of adverse academic outcomes. Proceedings of the ACM
SIGKDD International Conference on Knowledge Discovery and Data Mining, 2015-August, 1909–1918.
https://doi.org/10.1145/2783258.2788620
Letkovsky, S., Jencova, S., Vasanicova, P., Gavura, S., & Bacik, R. (2023). Predicting bankruptcy using artificial
intelligence: The case of the engineering industry. Economics and Sociology, 16(4), 178-190.
doi:10.14254/2071- 789X.2023/16-4/8
Li, B., Samsi, S., Gadepally, V., & Tiwari, D. (2022). Clover: Toward Sustainable AI with Carbon-Aware
Machine Learning Inference Service. Findings of the Association for Computational Linguistics: NAACL ,
1962–1970. https://doi.org/10.1145/3581784.3607034
Lison, P., Pilán, I., Sánchez, D., Batet, M., & Øvrelid, L. (2021). Anonymisation Models for Text Data: State of
the art, Challenges and Future Directions. ACL-IJCNLP 2021 - 59th Annual Meeting of the Association for
Computational Linguistics and the 11th International Joint Conference on Natural Language Processing,
Proceedings of the Conference, 4188–4203. https://doi.org/10.18653/V1/2021.ACL-LONG.323
McDonald, J., Li, B., Frey, N., Tiwari, D., Gadepally, V., & Samsi, S. (2022). Great Power, Great Responsibility:
Recommendations for Reducing Energy for Training Language Models. Findings of the Association for
Computational Linguistics: NAACL 2022 - Findings, 1962–1970.
https://doi.org/10.18653/v1/2022.findings-naacl.151
Molnar, C. (2019). Interpretable Machine Learning A Guide for Making Black Box Models Explainable. Leanpub.
http://leanpub.com/interpretable-machine-learning
Mullainathan, S., & Obermeyer, Z. (2017). Does Machine Learning Automate Moral Hazard and Error? American
Economic Review, 107(5), 476–480. https://doi.org/10.1257/AER.P20171084
Ohlenburg, T. (2020). AI in social protection -exploring opportunities and mitigaiting risks. Deutsche Gesellschaft
für Internationale Zusammenarbeit. https://socialprotection.org/discover/publications/ai-social-protection
Open AI. (2023). GPT-4 Technical Report. In Submitted.
Patsakis, C., & Lykousas, N. (2023). Man vs the machine: The Struggle for Effective Text Anonymisation in the
Age of Large Language Models. Preprint. https://arxiv.org/abs/2303.12429v1
Roshchyk, I., Oliinyk, O., Mishchuk, H., & Bilan, Y. (2022). IT Products, E-Commerce, and Growth: Analysis
of Links in Emerging Market. Transformations in Business & Economics, 21(1), 209-227.
Ross, C., & Swetlitz, I. (2017). IBM pitched Watson as a revolution in cancer care. It’s nowhere close. moz-
extension://b3f4dfef-8263-4daa-9eff-df6d2dfbba2b/enhanced-
reader.html?openApp&pdf=https%3A%2F%2Fwww.preventcancer.org%2Fwp-
content%2Fuploads%2F2018%2F06%2FIBM_pitched_Watson_as_a_revolution_in_cancer_care.pdf
Sculley, D., Holt, G., Golovin, D., Davydov, E., Phillips, T., Ebner, D., Chaudhary, V., Young, M., Crespo, J.-F.,
& Dennison, D. (2015). Hidden Technical Debt in Machine Learning Systems. In Advances in Neural
Information Processing Systems (Vol. 28).
Shevlane, T., Farquhar, S., Garfinkel, B., Phuong, M., Whittlestone, J., Leung, J., Kokotajlo, D., Marchal, N.,
Anderljung, M., Kolt, N., Ho, L., Siddarth, D., Avin, S., Hawkins, W., Kim, B., Gabriel, I., Bolina, V., Clark,
J., Bengio, Y., … Dafoe, A. (2023). Model evaluation for extreme risks. Preprint.
https://arxiv.org/abs/2305.15324v2
Seniutis, Gružauskas, Lileikiene, & Navickas
24
Strickland, E. (2019). IBM Watson, heal thyself: How IBM overpromised and underdelivered on AI health care.
IEEE Spectrum, 56(4), 24–31. https://doi.org/10.1109/MSPEC.2019.8678513
Suresh, H., & Guttag, J. (2021, October 5). A Framework for Understanding Sources of Harm throughout the
Machine Learning Life Cycle. ACM International Conference Proceeding Series.
https://doi.org/10.1145/3465416.3483305
Torralba, A., & Efros, A. (2011, May). Unbiased look at dataset bias. In CVPR 2011 (Pp. 1521-1528). IEEE.
https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=5995347
United Nations. (2015). Transforming our world : the 2030 Agenda for Sustainable Development.
https://www.refworld.org/docid/57b6e3e44.html
United Nations. (2021). Recommendation on the Ethics of Artificial Intelligence (Issue November).
https://doi.org/10.7551/mitpress/14102.003.0010
United Nations. (2022). United Nations Activities on Artificial Intelligence (AI). https://eur-lex.europa.eu/legal-
content/EN/TXT/HTML/?uri=CELEX:52018DC0237
United Nations. (2023). Building fertile ground for data science in Uganda. United Nations Global Pulse.
https://medium.com/un-global-pulse/building-fertile-ground-for-data-science-in-uganda-a950dfd3ca0b
White House. (2022). Blueprint for an AI Bill of Rights. Making automated systems work for the American
peopleTEMS WORK FOR THE AMERICAN PEOPLE. https://www.whitehouse.gov/ostp/ai-bill-of-rights
Zhou, B., Li, B., Qi, P., Liu, B., Di, S., Liu, J., Pei, J., & Yi, J. (2023). Trustworthy AI: From Principles to
Practices; Trustworthy AI: From Principles to Practices. ACM Computing Surveys, 55(9), 177.
https://doi.org/10.1145/3555803
Authors’ Note
All correspondence should be addressed to
Miroslavas Seniutis
Institute of Sociology and Social Work, Vilnius University, Lithuania
Human Technology ISSN 1795-6889
https://ht.csr-pub.eu
Copyright of Human Technology is the property of Centre of Sociological Research and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use.
