Governance in the Context of Artificial Intelligence in Developing Countries ()
1. Introduction
The universal consensus holds that governance is the set of processes, norms, institutions, and practices through which an organization, a state, or any social or economic system is managed and directed. It involves decision-making and the way in which these decisions are implemented and enforced (Robles Carrillo, 2020). This concept is applicable to diverse settings such as: 1) Public governance: Management and administration of a country or territory by the government and other institutions. 2) Corporate governance: Norms and structures that regulate the operation of a company or corporation. 3) Environmental governance: Policies and actions to manage natural resources and the environment. 4) Global governance: Cooperation between states and international organizations to solve global problems.
Following the above, it must be understood that governance is not only about formal authority, such as a government, but also about the participation of actors such as the private sector, civil society, and international organizations (Hinojosa, 2021).
On the other hand, Artificial Intelligence (AI) is a field of computer science that seeks to create systems capable of performing tasks that normally require human intelligence (Suarez Xavier, 2021). These tasks include learning, reasoning, problem-solving, perception, and decision-making. AI applications are widespread, such as: 1) Facial and voice recognition. 2) Medical diagnosis. 3) Process automation.
AI in governance is primarily applied to improve decision-making, administrative efficiency, and transparency in public and government management (Filgueiras, 2021). These are some of the ways it is being used today: 1) Data-driven decision-making. 2) Government automation and efficiency. 3) Security and law enforcement. 4) Transparency and the fight against corruption. 5) Citizen participation and digital democracy.
In this regard, the research by Peña (2021) describes the creation of the artificial intelligence governance framework in the European Union, based on a triad of ethical norms, standardization, and hard-law regulation. It first addresses some bottom-up and top-down initiatives that have influenced the EU response. It then examines the novel EU policy on artificial intelligence governance and the emerging support structures for building this framework. It also critically assesses the three instruments that underpin the EU’s response to the challenge of AI: the ethical framework aimed at promoting a global standard; the standardization process aimed at facilitating technical development and promoting high standards of consumer and user protection; and, finally, the unfinished hard-law regulatory framework, which requires decisive EU intervention to mitigate risks and ensure the harmonious development of the data market and AI in Europe (Peña, 2021). The contribution (Serna, 2021) also highlights that the development and integration of Artificial Intelligence (AI) at a cross-cutting level in public sector organizations, going beyond specific sector-specific initiatives, requires new capabilities. A review of different approaches that address this issue highlights the importance of data and, more specifically, its governance in public administrations. To delve deeper into this topic, the different dimensions of data governance are analyzed and five components for its development are identified: strategy, data architecture and infrastructure, organization (including structure and processes), talent management and professional competencies, and the organization’s relationship model with its environment (López, 2021). Through conceptual reflection and its application to a case study of Barcelona City Council, with contributions from different components, lessons learned are highlighted and proposals are formulated for each of them. The conclusions highlight the need for an integrated institutional reinforcement strategy that links the different components of data governance linked to the development of Artificial Intelligence in the public sector (Hinojosa & Acosta, 2021).
The purpose of this research is to define the conditions under which developing countries can implement AI in their governance systems. This work focuses on the following areas of results: 1) Review of related literature. 2) Bibliometric analysis of the research topic: governance in the context of artificial intelligence in developing countries. 3) The challenges, opportunities, and prospects of artificial intelligence in the context of governance in developing nations.
2. Materials and Methods
The Scopus academic research platform and EndNote bibliographic manager were used as materials in this scientific contribution. Scopus is a bibliographic database of abstracts and citations of scientific articles, managed by Elsevier. It is one of the largest and most recognized sources of academic information, used by researchers, universities, and institutions worldwide (Bravo, Bennia, Naji, Fellouah, & Báez, 2020). The Scopus platform is defined by the following characteristic features: 1) It contains more than 84 million records, including articles from scientific journals, conferences, and books in various disciplines. 2) It indexes more than 25,000 peer-reviewed journals from more than 7000 publishers. 3) It offers metrics such as the H-index of authors and journals, and the CiteScore, which measures the influence of publications. 4) It allows searching by author, institution, keywords, and citations received. 5) It helps track how and where articles are cited over time.
On the other hand, EndNote is a bibliographic reference management software developed by Clarivate Analytics. It is primarily used to organize, store, and cite references in academic and research documents (Bravo-Hidalgo & Baez-Hernandez, 2019). Its defining features are: 1) It allows you to import, organize, and store bibliographic references in a personal database. 2) It integrates with word processors such as Microsoft Word to insert citations and generate bibliographies in different formats (APA, MLA, Vancouver, etc.). 3) It allows you to share reference libraries with other researchers and collaborate. 4) It connects with academic databases (such as PubMed, Scopus, and Web of Science) to import references directly. 5) It automatically creates reference lists based on the chosen citation style. 6) It allows you to attach and annotate PDF files within the reference library. 7) It facilitates access to the library from different devices. This tool is widely used by researchers, students, and academics to efficiently manage references and avoid citation errors (Bravo Hidalgo, 2018).
Using the search criteria “governance AND artificial AND intelligence” in the title, abstract, and keywords of the contributions in Scopus, and limiting it to contributions focused on societies in third-world nations or developing countries, 854 scientific contributions were detected. These research papers are grouped into original articles, review articles, conference papers, book chapters, books, and other less representative publications.
The detected information was processed using the EndNote computational bibliographic analysis and management tool. Furthermore, this tool allowed for the generation of each of the bibliographic references for this research, under the conditions and editorial policies of the journal that published this work.
3. Results
3.1. Review of Related Literature
The authors’ research (Osakwe, Mutelo, & Shilamba, 2021) highlights that artificial intelligence (AI) is advancing rapidly, and its use and impact have the potential to improve several existing structural inefficiencies in the performance of government functions. This paper explored how artificial intelligence can impact governance in developing countries. The study applied a quantitative approach using online survey questionnaires. The research instrument was distributed by sharing links through emails, Facebook, WhatsApp, websites, and other social media to relevant government institutions. In the end, 3203 participants responded to the questionnaire, which took 97 days to collect. The data collected from the surveys were quantitatively analyzed to obtain descriptive statistics using the Statistical Package for the Social Sciences (SPSS) and presented through graphs and charts. The findings show that AI will have a positive impact on government activities. The researchers’ manuscript (Satyanarayana, Guruprasad, & Balakrishnan, 2019) points out that automation is the buzzword in developing countries like India and China, where there is a huge population. Although independence came more than 70 years ago and technology is changing drastically every minute, most systems follow traditional methods to complete work. Here, there are nothing, but services provided by government departments like Revenue, Municipal, MRO, Education, RTO, etc. Technology is changing drastically. For example, 10 years ago having cordless phones was a dream, but today 90% of the population uses smartphones. Television has become smart, washing machines have become smart, refrigerators have become smart. What is not smart is society now, but not the systems or services provided by government departments. There are few departments implementing automation or intelligent systems for their services, such as the railways, where platform tickets can be purchased using a kiosk located at the train station without waiting in “Q”. They have invested heavily in this, but 80% to 90% of the systems are not properly maintained. In this research, the main objective is to make each government service a smart enabler, either by using existing hardware or by making minor changes to existing hardware by implementing high-level artificial intelligence techniques. The main advantage of this research is that it will improve the quality of service provided and employability.
The article (Aloamaka & Omozue, 2024) highlights that the rapid advancement of artificial intelligence (AI) offers substantial opportunities and significant challenges, particularly about human rights in developing countries. This article explores the legal issues and moral conundrums raised by the growing influence of AI technology in sectors such as healthcare, education, and governance. In developing countries, AI implementation often occurs without robust legal frameworks, potentially infringing on privacy, autonomy, and equity. This study employs a doctrinal research methodology to analyze the effects of AI on fundamental human rights, using case studies to illustrate both benefits and risks. It examines the role of international human rights law and local legal systems in mitigating adverse impacts. The findings highlight the complex interplay between technological innovation and human rights protection and propose practical recommendations for policymakers, technologists, and legal practitioners. Ultimately, the goal is to foster an environment in which AI can contribute to sustainable development while upholding the dignity and rights of all people in developing countries. This research contributes to the scientific understanding of the ethical and legal implications of AI in the context of human rights.
The personal contribution of (Filgueiras, 2022) highlights that the development of AI technologies has been a priority for governments around the world, forcing actors to create strategies and policies to accelerate and implement AI in industry, markets, and governments. This article analyzes the causal mechanism between the institutional dynamics of political regimes and AI development. Specifically, it compares 30 developing countries to understand the political and governance dynamics that explain the outcomes obtained with AI policies. Delving into the relationship between politics and policies, this article reflects on how different institutional frameworks produce different outcomes in terms of AI development. The research was based on data from the Transformation Index (BTI) of the Bertelsmann Stiftung and the Organization for Economic Cooperation and Development (OECD), using fuzzy set quantitative comparative analysis (QCA) to analyze the cases. The findings suggest that authoritarian countries perform better in producing development outcomes from AI.
Furthermore, this author, in this other work (Filgueiras, 2023), indicates that, in many situations, governance challenges in Latin America are more specific and depend on creating a perspective on highly unequal social and political realities, heterogeneous in terms of state capacities, market differences, and perspectives on sociotechnical ventures. The application of artificial intelligence in the region has involved different governance challenges, including sustainability, economic development, health, and education, among others. Its advancement makes it an essential tool for addressing these challenges in developing countries. Latin American countries are making progress in building strategic policies to master artificial intelligence to respond to different development problems. However, the way these strategic policies are built occurs within their framework. On the one hand, Latin American countries face the risk of technological dependence, including problems with access to data, critical infrastructure for AI development, and technical and operational capabilities. On the other hand, their strategic policies express elements of a break from a colonial logic that can be reinforced by sociotechnical systems based on artificial intelligence. This article will work with this dual framework that informs the construction and application of artificial intelligence in Latin America.
This article (Gehl Sampath, 2021) seeks to position AI-based activities and related policy developments in the Global South considering a technology industry with desires for continued expansion and self-governance. The article progresses along three themes. First, using field interviews and related information (anonymized), the article classifies expanding AI activities across several core sectors by companies, governments, and other agencies in the Global South into three categories. It then reviews current AI policy priorities and positions in developing countries, showing how the introduction of AI has been shaped by considerations of digital cannibalization by large Northern firms at the expense of a full-fledged debate on privacy and data protection. By showing how AI exemplifies structural inequality in the Global South, the article identifies four critical starting points in developing countries that make the increasing use of AI a matter of concern. While many of the challenges posed by AI are to some extent universal, the article argues that its ramifications may be far worse in the Global South. Countering them will require policy frameworks that take privacy protection into account in a more responsible manner. It concludes with suggestions for a more nuanced debate on industrialization, privacy protection, and development.
The publication (Nwafor, 2024) states that biases, particularly gender biases, are common in Artificial Intelligence (AI) systems, generating harmful effects that reinforce existing negative gender stereotypes and prejudices. Although gender mainstreaming is a topical and fashionable issue in written discourse, it has not yet been thoroughly implemented in practice. While the clamor for AI regulation is common worldwide, most government policies on the topic do not adequately account for gender inequalities. In Africa, Egypt, Rwanda, and Mauritius are at the forefront of AI policy development. By exploring these three countries as case studies, employing a feminist approach and using the African Union Strategy for Gender Equality and the Empowerment of Women 2018-2028 as a methodological guide, this study conducts a comparative analysis of gender considerations in their policy approaches to AI. It found a disconnect between gender equality/responsiveness and these countries’ AI strategies, demonstrating that gender has not yet been incorporated into these policies. The study offers key recommendations that provide an opportunity for African countries to be innovative leaders in AI governance by developing even stronger policies compared to Western AI policies that do not adequately address gender.
The article (Quindimil, 2024) explores the impact of artificial intelligence (AI) on international trade, focusing on the challenges and opportunities for developing countries, particularly in Latin America and the Caribbean (LAC). As AI advances, it creates regulatory challenges and ethical concerns, especially regarding human rights and public morals, which can lead to trade frictions not addressed in current international agreements. AI provides significant benefits to countries, such as cost reductions and efficiencies, along with challenges related to data privacy, competition, and security. For developing countries, AI adoption offers both opportunities and obstacles, with infrastructure and human capital constraints being significant obstacles. The article suggests flexible regulatory approaches to attract investment, promote digital services exports, and facilitate cross-border data flows. While global trade discussions, particularly within the WTO, are addressing AI-related issues, regional agreements have made more progress. However, these agreements lack comprehensive coverage and effective dispute resolution. It is important for LAC countries to participate in international negotiations to ensure their interests are represented and help shape global digital governance standards, enabling them to fully benefit from AI.
The study (Trajkovski, 2024) suggests that the arrival of AI will revolutionize governance and public administration, presenting both opportunities and challenges. This document offers a roadmap for public agencies, detailing the steps from preparation to widespread AI implementation. It proposes a skills framework encompassing technical, ethical, legal, and managerial aspects, complemented by recommendations for ongoing training. Emphasizing an ethical and human-centered approach, the document aims to foster innovative and responsible governance. Collaboration is highlighted as vital to accelerating AI adoption and equipping administrators with tools to navigate this complex yet promising landscape. The document also addresses the equality and inclusion challenges posed by AI, particularly in bridging the gap between the Global North and the Global South, using international examples from both developed and developing countries. These insights ensure a comprehensive perspective on the integration of AI into public administration, promoting a holistic and nuanced approach to addressing these challenges.
The researcher (Truby, 2020) demonstrates that the unregulated release of experimental AI by Big Tech poses risks to the achievement of the United Nations Sustainable Development Goals (SDGs), with vulnerability for developing countries. The goal of financial inclusion is threatened by the imperfect and uncontrolled design and implementation of AI decision-making software that makes important financial decisions affecting customers. Automated decision-making algorithms have shown evidence of bias, lack of ethical governance, and limit transparency in the basis of their decisions, causing unfair outcomes and amplifying unequal access to finance. Poverty reduction and sustainable development goals are threatened by the potential exploitation of developing countries by Big Tech using AI to harvest data and profits. Stakeholder progress in preventing financial crime and corruption is further threatened by the potential misuse of AI. Considering these risks, Big Tech’s unscrupulous history means they cannot be trusted to operate without regulatory oversight. The article proposes effective preemptive regulatory options to minimize scenarios where AI harms the SDGs. This article analyzes internationally accepted AI governance principles and argues for their implementation as regulatory requirements governing AI developers and coders, with compliance verified through algorithmic audits. Furthermore, it argues that AI governance frameworks should mandate a benefit to the SDGs. The article argues that proactively predicting such problems can enable continued AI innovation through well-designed regulations that adhere to international principles. The risks of unregulated AI causing harm to human interests are highlighted, where a negative public and regulatory backlash can lead to overregulation that could harm otherwise beneficial AI development.
3.2. Bibliometric Analysis of the Research Topic: Governance in the Context of Artificial Intelligence in Developing Countries
This research topic, governance in the context of artificial intelligence in developing countries, is currently widely accepted by the international scientific community. Figure 1 shows the percentage distribution of scientific contributions by type. It can be observed that the highest percentage of research disseminated is in the form of articles, followed by research disseminated as conference papers. These results demonstrate the dynamism of scientific activity in this research topic. The conception and publication of a scientific article is faster than that of other types of scientific contributions, such as book chapters or books.
Figure 1. Distribución porcentual de los diversos tipos de documentos publicados en Scopus.
The countries with the highest scientific productivity in this line of research are India, by far, followed by the United States of America, the United Kingdom, Morocco, and China. Figure 2 shows the results obtained in the Scopus database, using the previously defined search criteria.
Figure 2. Number of scientific publications published by nations and registered in Scopus.
Figure 3. The increase in the number of scientific publications detected in Scopus over the last 20 years.
By observing the evolution of scientific productivity, detected in the Scopus academic research database and using established search criteria, the historical data shown in Figure 3 was compiled. It can be seen that, starting in 2019, a growing number of scientific contributions published in Scopus related to governance in the context of artificial intelligence in developing countries began.
4. Discussions
Governance in the context of artificial intelligence in developing countries presents challenges and opportunities that must be addressed to ensure that technological progress is inclusive, ethical, and beneficial to all sectors of society (Abdelati Nasr, Ebed Emam, Adam Ahmed, & Alshamrani, 2024). This section of the document details the challenges, opportunities, and prospects of artificial intelligence in the context of governance in developing nations (Abdulkareem, 2024).
Inequality in access to technology: Developing countries face economic barriers, poor infrastructure, and limited human resources trained in AI. This can lead to technological gaps, where only certain sectors or elites benefit from AI, while most of the population is left behind (Abdelbary & Shreaf, 2024).
Lack of clear regulatory frameworks: AI governance requires strong legal and regulatory frameworks. In many developing countries, there are still no clear regulations on how to use AI ethically, safely, and effectively. This can lead to irresponsible use of technology (Justo-Hanani, 2022).
Privacy and security concerns: AI relies on large amounts of data, raising serious privacy concerns. In developing countries, where data protection infrastructure is often weak, AI could be exploited without people’s proper consent or the necessary security measures (Silva, 2022).
Insufficient training and education: The lack of adequate AI training in both the public and private sectors can hinder the development of effective policies. Without building local AI capacity, developing countries may rely too heavily on imported models that are not always adapted to their socioeconomic realities (Rana, Chatterjee, Dwivedi, & Akter, 2022).
Boosting economic development: AI has the potential to transform various sectors, such as agriculture, healthcare, education, and infrastructure, which could improve economic and social well-being. Appropriate AI governance can ensure that these advances are distributed equitably.
Improving public decision-making: AI can be a valuable tool for governments in analyzing data and improving decision-making. With proper regulation, developing countries could use AI to improve public services, optimize resource use, and combat corruption (Escobedo et al., 2024).
International collaboration: Developing countries can benefit from international AI governance that allows for the sharing of knowledge, resources, and best practices. Furthermore, cooperation with international organizations and the private sector can facilitate access to and implementation of AI technologies.
Ethics and human rights: AI governance can be an opportunity to promote ethical standards globally. In developing countries, integrating ethical principles from the outset could ensure that AI is used for the common good, respecting human rights and promoting social equity (Awashreh & Ramachandran, 2024).
Development of national regulatory frameworks: Developing countries must work to create national policies that clearly regulate the use of AI, with an emphasis on ethics, privacy, and inclusion. It is also essential that these policies be flexible to adapt to rapid technological advances.
Promote education and training: It is essential to promote AI education and training at all levels. This includes both technical training for engineers and scientists, as well as training for public officials and policymakers to understand how AI can benefit or harm society (Almaqtari, 2024).
Promote public-private collaboration: Governments should work hand in hand with the private sector, NGOs, and universities to create technological solutions tailored to local realities. This may include initiatives to improve technological infrastructure or provide equitable access to the benefits of AI.
Ensure citizen participation: Involving society in the creation and implementation of AI policies is crucial to ensure that technological decisions are inclusive and represent the needs of the general population. This includes holding public consultations and incorporating diverse voices in decision-making (Saxena et al., 2023).
Foster international cooperation: AI is a global phenomenon, and its governance should also be international. Developing countries can collaborate with international organizations, such as the UN, to develop global standards that help mitigate risks and promote the equitable development of technology.
Governance and artificial intelligence (AI) in developing countries is a complex issue that involves a series of challenges and opportunities. In these countries, the implementation of AI can bring improvements in key areas such as health, education, agriculture, security, and public management. However, there are also risks and barriers that must be addressed to ensure that its implementation is effective, inclusive and ethical (Xia, Semirumi, & Rezaei, 2023).
5. Conclusion
The development and application of AI in the governance environment poses benefits, but also ethical and social challenges, such as algorithmic bias, privacy, and the impact on employment, among others. AI governance in developing countries is a complex issue that involves the creation of balanced policies, the promotion of technological education, and international cooperation. If managed correctly, developing countries have the potential to use AI to transform their economies and improve the lives of their citizens, provided that inclusive, ethical, and locally adapted governance frameworks are implemented. In other words, artificial intelligence offers great potential for developing countries, but its success will depend on adequate governance, the strengthening of technological infrastructure, and an ethical and equitable approach that addresses both the associated opportunities and risks.