CHAPTER TWO: LITERATURE REVIEW

Introduction

AI has changed the way we live and work, bringing many new opportunities but also
presenting several challenges.25 AI technology can address key issues related to climate
change, energy, agriculture, healthcare, intellectual property, ICT, finance, governance, legal,
and business processes.26 Despite its potential, there are concerns that AI deployment may
negatively impact society, raising ethical questions. Among the negative outcomes, AI
technology can carry gender and racial biases, misinformation, echo chambers, fake news, and
27
hate speech. Additionally, AI can aggravate social, political, and economic inequalities,
particularly affecting persons with disabilities, the elderly, children, and women, and further
widening the digital divide. 28 Also, deploying AI can lead to human rights abuses while raising
serious concerns about data privacy, security, and cross-border data flows.29

As AI continues to transform various sectors globally, including healthcare, finance, and public
administration, it is crucial to establish strong legal frameworks that can effectively govern its
deployment. This chapter aims to provide a comprehensive review of existing literature on AI
regulation, focusing on the ethical and legal challenges associated with its deployment. 30
Global Overview of AI Regulation

Artificial Intelligence (AI) regulation is an evolving discipline that addresses the need to
balance technological advancement with ethical concerns, privacy, and societal impact. As AI
technologies rapidly advance and become more integrated into daily life, governments and
organizations worldwide are developing regulations to ensure these systems are
used In reviewing the regulatory challenges surrounding Artificial Intelligence (AI),
particularly in Ethiopia, a notable scarcity of documentation addressing these issues was
observed. While global AI deployment has generated extensive legal and ethical debates,
research on these matters within the Ethiopian context remains limited. Consequently, this
review examines the current AI regulatory landscape both globally and in developing
countries, with a focus on the few studies and documents that pertain specifically to Ethiopia.
In our nation Ethiopia, numerous initiatives are underway to promote the development and
adoption of AI technology. Key actions taken include establishing dedicated institutions,
enhancing human capital in the sector, assessing the sector’s capacity for growth and benefit,
preparing implementation plans, and conducting research. However, further steps are needed to
build a competitive capacity that ensures AI technology is used broadly across social, economic,
and political domains, while also safeguarding national security. Developing and implementing
a comprehensive policy to guide the sector is one such step. The creation of this policy will
contribute significantly to achieving sustainable development goals at both national and
global levels
At the global level, scholars such as Binns and Crawford have emphasized the importance of
establishing a robust regulatory framework to address key ethical concerns, including bias,
transparency, and accountability in AI systems. 18In particular, Finoc chiaro discusses the need
for a detailed regulatory approach to AI that balances innovation with ethical considerations,
highlighting that effective regulation must adapt to the rapidly evolving landscape of AI
technologies.19 Similarly, O'Neil emphasizes the dangers of algorithmic bias, demonstrating
how poorly regulated AI can perpetuate systemic inequalities, particularly against marginalized
communities.20
In the African context, the deployment of Artificial Intelligence presents significant
opportunities alongside critical challenges. Arakpogun et al. highlight that while AI has the

potential to drive economic growth and innovation in Africa, it also faces substantial
hurdles related to infrastructure, regulatory frameworks, and workforce preparedness. 21
Another author Jaldi emphasizes the economic benefits of AI, noting the necessity of
coherent legal frameworks to address pressing legal challenges, particularly regarding data
protection and privacy rights.22

Additionally, regarding regulatory approaches, Gaffley et al. advocate for a human rights–
based framework for AI regulation, emphasizing the need for rules that prioritize individual
rights and ethical considerations, particularly for marginalized communities. Adams and
Shyllon further contribute to the discourse by highlighting AI‘s ethical and human rights
dimensions, arguing that AI deployment must be aligned with human rights principles and
broader social values equity.23

Turning to Ethiopia, there is a significant gap in research on the governance of Artificial

Intelligence, particularly concerning ethical and regulatory frameworks. A notable contribution
to this discourse is the study by Jima, Tarekegn, and Debelee, which examines AI policy,
ethics, and privacy rights within Ethiopia their analysis highlights the critical need for privacy
safeguards and ethical guidance in AI deployment, particularly distinguishing between
narrowly focused data protection laws and broader normative frameworks for responsible AI.
While their study emphasizes the urgency of establishing such measures, this research expands
the discussion by systematically assessing Ethiopia‘s current legal gaps, ethical challenges, and
international best practices to propose a comprehensive, contextually relevant regulatory
approach.
The authors also draw on international examples, such as the General Data Protection
Regulation (GDPR) and the California Consumer Privacy Act (CCPA), to illustrate different
approaches to privacy governance. Additionally, they examine economic considerations, noting
that while robust data governance frameworks can catalyze innovation, they may also create
competitive disparities, particularly disadvantaging smaller firms.
Unlike previous studies that focus primarily on AI skills, infrastructure, and research and
development, this research addresses the overlooked legal and ethical challenges in Ethiopia‘s
AI governance. It examines regulatory gaps, accountability issues, and societal risks,
proposing

solutions to ensure responsible and rights-respecting AI deployment. In doing so, it

complements existing technical and policy-focused research while filling a critical gap in
understanding the legal and ethical dimensions of AI adoption in Ethiopia.

ARTIFICIAL INTELLIGENCE POLICY OF ETHIOPIA

The development and application of AI in our nation are still in their infancy, with significant
challenges stemming from the lack of robust data infrastructure and high- quality data. This
limitation has adversely affected the training and optimization of AI models. Furthermore, the
limited AI infrastructure and inadequate internet connectivity in our nation have hindered the
widespread implementation of AI solutions. Start-ups focusing on AI technology face numerous
challenges, including the early stage of AI platform development, a lack of financial and
technical support, and the absence of other enabling conditions. Additionally, the shortage of
skilled professionals in the AI field has significantly impacted research and development, as
well as the growth of creativity and innovation within the sector.
The absence of a regulatory framework for AI development and application, coupled with
insufficient capacity to enforce such regulations, has led to gaps in areas such as data privacy,
ethical AI use, and legal accountability. These challenges have further slowed down AI adoption
and growth. To guide thegovernment in addressing these issues, it is crucial to establish a
comprehensive

AI platform and leverage the technology’s potential for national growth. Considering these
challenges, several key actions can serve as solutions to foster AI development. First, investing
in data collection, storage, and management systems to develop an integrated data infrastructure
is essential, alongside promoting collaboration to gather high-quality data. Second, resources
must be allocated to enhance AI infrastructure and internet connectivity to improve nationwide
accessibility to AI technologies. Third, supporting

AI-focused start-ups can be achieved by developing incubation centers, organizing advisory

programs, and fostering collaboration among academia, industry, private enterprises and the
government. Fourth, addressing the need for skilled personnel in the AI sector requires
investment in education and training programs to cultivate relevant expertise. Finally, ensuring
accountable and sustainable AI growth necessitates tackling issues of data privacy, ethics, and
accountability through the formulation and implementation of a comprehensive AI regulatory
framework. By collectively implementing these measures, we can create an enabling
environment for AI innovation and adoption in our nation.

THE NEED FOR THE POLICY, PRINCIPLES, AND OBJECTIVES

The adoption of a national AI policy presents significant benefits. Crafting an AI policy
establishes the direction necessary to derive effective benefits from the technology across areas
such as innovation, agricultural productivity, education, healthcare, service delivery, and other
key sectors. Such a policy plays a critical role in creating a framework to mitigate risks and
challenges associated with AI, including biases in AI products and services, discrimination, data
privacy breaches, security threats, and adverse environmental impacts. Moreover, the policy
will serve as a foundational document that ensures AI products and services developed, adapted,
and implemented in the country align with human rights, societal values, and uphold both rights
and responsibilities. By defining the relationships and partnerships between government bodies,
private institutions, educational and research entities, and other stakeholders, the policy will
facilitate the implementation of AI and enhance international engagement through a structured
communication framework.

The Global AI Index Report assesses and reports on the progress of countries in AI
development, ranking them annually based on their achievements. According to this report,
Ethiopia currently ranks quite low. To improve this ranking, one crucial step is to develop a
comprehensive policy framework to govern AI effectively. Such a policy, by establishing clear
guidance for the AI sector, will significantly contribute to improving the country’s standing in
the report. Additionally, it will lay the foundation for global cooperation and partnerships within
the AI sector.

As a sovereign nation, Ethiopia requires a policy that addresses critical issues and
implementation strategies necessary for fostering a robust AI development, governance, and
administration system. This system aims to support the country’s broader development,
democratization, and security efforts, ultimately benefiting its citizens. Thus, this policy will be
instrumental in advancing Ethiopia’s goals in the AI sector, while also serving as a roadmap for
achieving the digital Ethiopia strategy by 2025.

Objective of AI OBJECTIVES of AI policy

This policy shall have the following specific objectives:

• Promoting the culture of using AI technology;
• Making the resource obtained from AI contribute to the overall national production and
creating the capacity that makes Ethiopia a destination for AI development in Africa;
• Building conducive environment for AI development and growth;
• Leading the development of the sector by establishing necessary monitoring system;
• Encouraging job creation and the use of AI research and development for problem- solving;
• Promoting participation of the private sector in AI development and also encouraging
beginners;
• Creating the capacity to supply AI products and services to local and external markets.
BASIC POLICY ISSUES AND DIRECTIONS
HUMAN RESOURCE DEVELOPMENT
To effectively harness AI technology in our nation, it is crucial to cultivate a skilled human
resource base capable of its adoption, utilization, and development. Given the novelty of AI in
our society, there exists a significant awareness gap. Additionally, challenges related to
knowledge, skills, and attitudes necessary for the effective use and advancement of AI are
apparent.
To bridge these gaps and overcome these challenges, the following policy directives will be
implemented
AI INFRASTRUCTURE DEVELOPMENT
Countries that have excelled in AI have successfully established the necessary infrastructure,
largely by integrating the technology within private sector institutions and gradually building
advanced computational capabilities. In some cases, progress in AI was achieved primarily due
to significant government investments aimed at developing the required technological
infrastructure.
To foster AI growth, it is essential to leverage existing information technology infrastructure
through collaboration with educational, research and technology institutions. It is also important
to develop shared infrastructure, especially those requiring substantial resources, by partnering
with other entities or by securing them through cooperative initiatives and donations.
Competitiveness in the AI sector can be further enhanced by fostering relationships and
partnerships with developed countries that have extensive experience in AI, and by swiftly
addressing the challenges that arise in this area. Therefore, to expand AI infrastructure and
bridge existing gaps, the following policy measures should be implemented.

DATA MANAGEMENT
Data accessibility is fundamental for the growth and advancement of AI. Consequently, high-
quality data curated by government agencies, private institutions, and individuals has played a
crucial role in the successes observed in countries with thriving AI sectors.
However, in our nation, there are significant challenges hindering the development of the AI
sector. These include limited awareness regarding data handling and utilization, insufficient
availability of organized digital data, inadequate analysis of existing data, the absence of a data-
sharing system, and the lack of an inclusive data collection methodology.
To foster AI development, it is imperative for all government and private institutions, as well as
different segments of society and individuals, to adopt standardized practices for data collection,
cleansing, sharing, and management.
The government must take responsibility for ensuring the availability of quality data by
establishing a legal framework for data development, use, and management; building
implementation capacity; issuing relevant standards; and creating supportive administrative
structures. In this regard, the following policy directions should be implemented to address the
challenges and support AI growth

RESEARCH AND DEVELOPMENT

Globally, the AI sector is experiencing rapid growth, with new innovations emerging across
various institutions and countries to modernize and simplify human life. Consequently, many
nations are investing significant resources to enhance their research and development
capabilities within the AI sector.
In our nation, however, the research ecosystem supporting AI development and innovation
remains in its infancy. To foster effective use of AI research and innovation for development, a
robust system must be in place to support research and development initiatives. AI research in
Ethiopia should be geared primarily toward addressing social and economic challenges,
contributing to the achievement of national development goals. Additionally, it is crucial to
strengthen the capacity of research institutions to facilitate focused AI development and
innovation. This requires the establishment of a research platform that continuously fosters a
conducive environment for AI research and innovation.
Furthermore, research initiatives must be conducted collaboratively among industries, higher
education institutions, and research organizations. The primary objective of this policy is to
address existing challenges, ensuring the success of research endeavors. Therefore, to advance
AI research and development and bridge existing gaps, the following policy directions must be
implemented:

LEGAL FRAMEWORKS AND PROCEDURES TO IMPLEMENT THE POLICY

Following the expansion of the AI technology, various legal and ethical violations are observed
in relation to individual rights, data protection, religion and culture, and other similar key
societal vlaues. As a result, it is essnetial to prepare uniform and all inclusive legal frameworks
and ethical principles in a participatory, transparent, fair and sustainabnle manner to avoid the
risks associated with the implementation of the AI technology mainly in relation to data
security, administration, and monitoring systems.

LEGAL AND REGULATORY FRAMEWORKS FOR ARTIFICIAL

INTELLIGENCE GLOBAL AND ETHIOPIAN PERSPECTIVES
responsibly, transparently, and in alignment with fundamental rights responsibly. 31These
regulations aim to mitigate risks such as bias, privacy violations, and misuse while promoting
innovation.32Different regions are adopting varied approaches, from comprehensive legal
frameworks in the European Union to principles-based guidelines in the United Kingdom,
reflecting diverse priorities and regulatory philosophies. Below, I will state some of the major
regulations related to AI worldwide:

The European Union Artificial Intelligence (AI) Act

The EU AI Act is the world‘s first comprehensive AI law, the European Parliament's articles on
the EU AI Act provide a thorough overview of this pioneering regulation aimed at governing
artificial intelligence within the European Union.33 The Act introduces a risk-based classification
system for AI, categorizing systems from minimal to high risk, with stringent requirements
imposed on higher-risk applications such as those used in critical infrastructure or law
enforcement.34 The Act emphasizes transparency and accountability, mandating disclosure when
AI is in use and enforcing rigorous compliance checks for high-risk systems. Also, the Act is
designed to balance regulation with innovation, incorporating regulatory sandboxes for testing
AI systems under oversight. It also establishes significant penalties for non-compliance. 35 The
EU AI Act is positioned to set a global standard for AI regulation, potentially influencing other
jurisdictions.36
USA

Despite its leading role in AI research and technology, the United States, as of January 2024,
still lacks a binding federal law that specifically regulates the development, deployment, and use
of AI.37This shows the fragmented nature of the country‘s approach to AI regulation, resulting in
a complex collection of laws and non-binding frameworks.38Although the country has some AI-
related legislation, these laws primarily address specific administrative issues
within the federal government and are largely irrelevant beyond the public sector. The
regulatory environment is instead characterized by a fragmented mix of existing laws that
may address certain AI-related risks, such as privacy, discrimination, and intellectual
property theft, on a case-by-case basis. 39 However, the USA lacks comprehensive regulation
specifically targeting AI technology itself, with existing laws covering aspects such as the
California Consumer Privacy Act and the Digital Millennium Copyright Act. 40

China

China is at the forefront of establishing some of the world‘s earliest and most detailed
regulations for artificial intelligence (AI), including measures for recommendation algorithms,
41
synthetically generated images, and chat bots similar to Catgut. These regulations are set to
significantly influence how AI technology is developed and deployed within China and on a
global scale, affecting both Chinese technology exports and international AI research networks.
Key regulations include the 2021 rules on recommendation algorithms, which address issues
like excessive price discrimination and algorithmic scheduling, the 2022 regulations on deep
synthesis, which mandate clear labeling of synthetic content and the 2023 draft rules for
generative AI, which impose stringent requirements on data accuracy and model outputs. 42 All
three regulations emphasize information control, requiring developers to file with China‘s new
algorithm registry and pass security self-assessments.43 While Western perspectives may
dismiss these regulations as geopolitically motivated, they offer valuable insights into different
regulatory approaches and their potential impacts on global AI governance.44

AI Regulatory Approaches

AI regulatory approaches vary widely but generally fall into several key categories. The
risk- based approach classifies AI systems based on their potential impact, with higher-risk
applications subject to more stringent regulations. 45 This approach is flexible and promotes
innovation by focusing regulatory efforts on the most critical applications. Sector-specific
regulations tailor rules to the needs of particular industries, such as healthcare or finance,
providingdetailedoversightrelevanttoeachsector'[Link]-basedregulations focus
on broad ethical guidelines and principles, allowing for a flexible interpretation and
application across different contexts. 46 Outcome-based regulations emphasize the results or
effects of AI systems rather than specific processes, aiming to ensure that AI technologies
achieve desired outcomes while allowing room for innovation. 47 Each approach has its
strengths and limitations, and a combination of these methods can offer a more balanced
framework for addressing the diverse challenges posed by AI technologies.
The Role of International Organizations in Setting AI Standards
International organizations play a crucial role in setting standards for artificial intelligence (AI)
by developing guidelines, encouraging global cooperation, and establishing frameworks that
ensure ethical and effective AI deployment. .48

These organizations, such as the International Organization for Standardization, 49 the

International Telecommunication Union,50 and the Organization for Economic Co-operation
and Development,51 work to create universally accepted standards that address safety, privacy,
and fairness in AI systems. By bringing together experts, policymakers, and stakeholders from
various countries, these entities help Harmonize regulations, promote best practices, and
mitigate risks associated with AI.

Technology their efforts contribute to building trust in AI, ensuring that its development
aligns with global ethical norms and supports sustainable innovation. 52

Ethical Challenges of AI Deployment

AI in the Court of Law

The deployment of AI in judicial systems, often referred to as the "automation of justice,"

offers the promise of greater efficiency and objectivity in legal decision-making.53 However,
significant ethical challenges arise, particularly concerning transparency, bias, accuracy, and
privacy. AI systems used in courts often function as black boxes, with their decision-making
processes not easily interpretable by judges, lawyers, or defendants. This lack of transparency
can undermine trust in judicial outcomes and raise concerns about accountability, particularly
when biases embedded in AI algorithms perpetuate existing societal inequalities. 54

The COMPAS algorithm is one of the most cited examples of biased AI in the criminal justice
system. This risk assessment tool was designed to predict the likelihood of a defendant
reoffending and is often used by judges during sentencing to guide decisions. However, a 2016
investigation by Republican found that COMPAS was significantly biased against Black
defendants. The study showed that Black defendants were almost twice as likely to be
incorrectly classified as high-risk compared to white defendants. In contrast, white defendants
were often mislabeled as low-risk despite having similar criminal histories.
The ethical dilemma here centers on the algorithm's lack of transparency and its impact on
fairness and justice. COMPAS operates as a "black box," making it impossible for
defendants and their attorneys to understand or challenge the basis of its risk scores. This
raises critical concerns about due process, accountability, and the right to a fair
trial.55Questions about accountability, fairness, and the right to due process, showing how
biased algorithms can perpetuate systemic discrimination in the judicial system 56

Privacy Concerns in Healthcare: Google’s Project Nightingale

Project Nightingale was a controversial collaboration between Google and Ascension, one of
the largest healthcare providers in the United States. 57 Launched in 2019, this project involved
the transfer of millions of patients' medical records to Google without the patients' knowledge or
explicit consent.58 The goal was to leverage AI and machine learning to improve healthcare
delivery and patient outcomes. However, the partnership raised significant ethical concerns
regarding privacy, data security, and consent.59

Critics argued that patients were not informed about how their sensitive medical data was
being used, violating their privacy rights. There were also concerns about how Google might
use this data beyond healthcare purposes, potentially exposing patients to data breaches or
misuse of their information.60 This case clearly shows the ethical challenges of using AI in
healthcare, particularly the need for transparency, informed consent, and stringent data
protection measures to safeguard patient privacy.
Bias in AI
AI systems often mirror societal biases present in the data they are trained on. The prevalence of
gender and racial biases in AI, particularly in image and search algorithms, demonstrates how AI
can reinforce harmful stereotypes.61 For instance, biased search results that predominantly
feature men when searching for ―greatest leaders‖ reflect the systemic undervaluation of
women's contributions. Searches for ―greatest leaders‖ predominantly feature men, while
―schoolgirl‖ images frequently include sexualized costumes, unlike ―schoolboy‖ images. 62 These
biases not only perpetuate existing stereotypes but also influence how individuals are perceived
and judged in various contexts, including hiring, education, and online representation
education, and law enforcement.63 To mitigate these biases, it is essential to develop strategies
for creating more inclusive datasets and refining AI algorithms to ensure fairer and more
equitable outcomes globally.

Facial recognition technology has also been increasingly adopted by law enforcement agencies
worldwide. However, these systems often exhibit racial and gender biases, leading to wrongful
identifications and arrests. A high-profile case illustrating this ethical dilemma occurred in
Detroit, Michigan, where Robert Williams, an African American man, was wrongfully arrested
based on a faulty match from a facial recognition system. 64 The software had mistakenly
identified Williams as a suspect in a theft, despite clear discrepancies between his appearance
and the actual suspect.65This case exemplifies the ethical issues surrounding the use of AI in
law enforcement, particularly the dangers of relying on biased technologies that
disproportionately affect marginalized communities. It also raises questions about
accountability when AI-driven errors have real-world consequences, such as wrongful arrests
and potential violations of civil rights.66
D Bias in Healthcare AI: The Optum Algorithm Case
An AI algorithm developed by Optum, used to identify patients who would benefit most from
high-risk care management programs, was found to exhibit racial bias.67A study revealed that the
algorithm systematically assigned lower risk scores to Black patients compared to white patients
with similar health conditions.68 This discrepancy arose because the algorithm used healthcare
costs as a proxy for health needs, unintentionally prioritizing patients who historically had
greater access to healthcare services typically favoring white patients. 69
E. AI in Art
The use of AI in creating art challenges traditional notions of authorship, creativity, and
copyright. AI-generated works, such as paintings mimicking Rembrandt's style or the creation of
entirely new artistic pieces through generative models, raise complex questions about
originality, intellectual property rights, and the role of human creativity in the artistic process.70
Completion of Schubert‘s Symphony No. 8, complicate questions of intellectual
property and

it raises ethical concerns about the value of human creativity and the potential exploitation of
original artists' styles without proper attribution or compensation. Furthermore, these
developments necessitate new legal frameworks to address originality, plagiarism, and
ownership, especially as AI-created art becomes more prevalent in the global market. 72
Ethical and Legal Challenges of AI Deployment in Africa
Africa faces double legal challenges in AI governance due to various factors, including
regulatory gaps, infrastructural deficits, socio-economic disparities, and ethical concerns. While
AI holds significant potential for driving economic growth and addressing social issues on the
continent, its governance remains fraught with challenges that require urgent attention to ensure
equitable and responsible deployment. Below is an overview of the specific legal challenges
Africans face in AI governance.
Lack of Comprehensive AI Regulation

One of the most pressing challenges in AI governance in Africa is the absence of

comprehensive regulatory frameworks specifically tailored to AI technologies. While some
African countries, such as Egypt and Rwanda, have developed national AI strategies, most
nations lack specific AI laws or regulations to govern the ethical use, development, and
deployment of AI systems.73This regulatory gap makes it difficult to address critical issues
such as data protection, algorithmic transparency, and accountability, leaving citizens
vulnerable to the unintended consequences of AI, including bias, discrimination, and
privacy violations.
Data Privacy and Protection Challenges

Data privacy and protection remain significant legal challenges for AI governance in Africa.
ManyAfricancountrieshaveeitheroutdateddataprotectionlawsorlackthemaltogether, making it
difficult to regulate AI systems that rely on large datasets, often including sensitive personal
information74 For instance, while the African Union adopted the Malabo Convention
on

Cyber Security and Personal Data Protection in 2014, only a few countries have ratified it,
limiting its effectiveness in creating a unified data protection standard across the
[Link],Africancitizensfaceincreasedrisksofdatabreac
hes, misuse of personal data, and insufficient safeguards against surveillance.
Bias and Discrimination in AI Algorithms

Bias in AI algorithms presents a significant legal and ethical challenge in African contexts. AI
systems trained on datasets that do not adequately represent African populations can produce
biased outcomes, reinforcing existing social and economic inequalities. 75 For example, facial
recognition technologies have been shown to have higher error rates for people with darker skin
tones, leading to misidentifications that can have serious legal implications. 76 The lack of legal
requirements for auditing and mitigating bias in AI systems further exacerbates this issue,
making it challenging to hold developers accountable for discriminatory practices.
Inadequate Legal Frameworks for Intellectual Property and Innovation
African innovators and entrepreneurs face legal challenges related to intellectual property (IP)
protection, particularly in the AI sector.77 Existing IP laws in many African countries are not
well-suited to address the complexities of AI, such as ownership of AI-generated content and the
protection of AI-driven innovations.78 The lack of clear legal guidance on IP rights in the context
of AI stifles innovation, as creators may be reluctant to invest in AI development without
assurances that their work will be adequately protected and rewarded.
Cross-Border Data Governance and Sovereignty Issues
The global nature of AI technologies presents challenges for African countries in managing
cross-border data flows and maintaining data sovereignty. AI systems often require data from
multiple jurisdictions, complicating regulatory oversight and raising concerns about how data
is collected, stored, and used across borders. African countries face the additional challenge of
negotiating data governance terms with powerful multinational tech companies that may
prioritize profit over compliance with local laws.79 The lack of harmonized cross-border data
regulations undermines the ability of African nations to protect their citizens' data and assert
control over their digital economies.
Limited Access to Legal and Technical Expertise

The shortage of legal and technical expertise in AI governance is a significant barrier to

effective regulation in Africa. Many African regulators lack the technical knowledge required
to understand AI systems, assess their impact, and develop appropriate legal responses. This
knowledge gap extends to legal practitioners, who may be ill-equipped to handle AI-related
cases due to the rapid evolution of AI technologies and the lack of specialized training. 80 This
limited capacity hinders the development of robust AI governance frameworks and leaves legal
systems struggling to keep pace with technological advancements.
The AI Legal Framework and Ethics in Ethiopia
Ethiopia‘s Council of Ministers has unanimously approved the AI policy in its 37th regular
meeting.81 Before the approval of the draft policy, there was no policy or law regulating AI. This
decision represents a major milestone for Ethiopia, advancing the nation's journey toward AI-
driven innovation to enhance socioeconomic development and strengthen technological
capabilities. When looking at the overall adoption of Ethiopia‘s AI policy, strategies, and
regulatory framework, it is clear that the country is taking initial steps to address both ethical
and legal considerations in the deployment of AI technologies. 82

Laws, progress remains limited compared to the rapid AI advancements seen in startups. Furthermore,
other AI stakeholders, such as universities, have begun identifying research priorities and are planning
to focus more on advancing AI development.83While new startups are already implementing AI and
other stakeholders are advancing research on AI, Ethiopia's laws, policies, and strategies are still in their
infancy. We currently have only one policy law, which has not yet been fully implemented.84

Current laws, such as the Cybercrime Proclamation No. 958/2016, touch upon digital security
and internet use but lack specific AI-related provisions. Significant gaps exist in areas like data
protection, privacy, and accountability. Ethiopia lacks comprehensive data protection laws

comparable to the GDPR in Europe, leaving questions about user data privacy and AI
accountability unresolved.
Soon, Ethiopia will face several ethical issues regarding AI, similar to other countries. The
main ethical challenges include the effects of AI on privacy, accountability, copyright,
transparency, bias and discrimination, and potential job loss, among other concerns. Such
problems show that there is an urgent need for a strong pillar of ethics so that society can
safely benefit from this technology. Other ethical challenges that Ethiopia may face, alongside
the rest of Africa, include low digital literacy, limited access to technology, and pronounced
socio-economic inequalities. These factors complicate the equitable deployment and regulation
of AI, potentially deepening existing societal divides and making it more difficult to ensure
that the benefits of AI are shared fairly across all segments of the population.
International Best Practices and Their Relevance to Ethiopia
When examining international best practices in AI regulation, several prominent frameworks
from global leaders provide valuable insights for Ethiopia the EU has been at the forefront with
its proposed Artificial Intelligence Act, which offers a comprehensive, risk-based classification
system for AI technologies.85 This system categorizes AI into four levels of risk—unacceptable,
high, limited, and minimal each with specific regulatory requirements. Additionally, the Act
emphasizes transparency and accountability, requiring AI providers to ensure responsible
deployment and reporting mechanisms. Disclose key information about their systems, and
integrates with the GDPR to ensure data privacy.86

In contrast, the United States follows a more decentralized approach, with sector-specific
regulations and ethical guide lines issued by agencies such as the FTC and the National
Institute of Standards and Technology (NIST). 87 this method allows flexibility but lacks a
unified framework, focusing on transparency, accountability, and risk management in fields
like finance, healthcare, and autonomous vehicles.

China, meanwhile, emphasizes strong state control over AI development, with a focus on
aligning AI advancements with national priorities. China's regulatory approach includes strict
Guidelines on algorithmic transparency, requiring companies to disclose algorithms that
shape public opinion.

Analysis of How These Best Practices Can Inform AI Governance in Ethiopia

The relevance of EU best practices can be seen in the potential for Ethiopia to adopt a risk-
based approach similar to the EU AI Act. By categorizing AI systems by risk levels, Ethiopia
can prioritize regulatory efforts and allocate resources more effectively. Additionally,
emphasizing transparency and accountability aligns with global trends toward ethical AI.
Implementing similar measures would help build trust in AI systems and ensure they operate
fairly and ethically.

Regarding U.S. best practices, Ethiopia could benefit from sector-specific guidelines to address
AI applications in critical areas such as healthcare, finance, and public safety. This approach
allows for tailored regulations that address the specific risks and needs of each sector.
Furthermore, adopting ethical guidelines, as suggested by the FTC and NIST, would ensure
that AI technologies in Ethiopia are developed and used responsibly.

While Ethiopia may not replicate China's state-controlled approach to AI development, it can
still learn from China‘s emphasis on aligning AI development with national priorities and
socioeconomic goals, thereby fostering strategic innovation while maintaining ethical
oversight.

The Adaptability of These Practices to Ethiopia’s Socio-Economic and Legal Context

From a socio-economic perspective, Ethiopia, like many developing countries, faces resource
constraints. Adapting international best practices for AI governance will require consideration
of these limitations, with a focus on scalable and cost-effective solutions. Additionally,
Ethiopia‘s focus on economic development and digital transformation necessitates a balanced
approach that fosters innovation while ensuring responsible AI use.

In terms of the legal framework, Ethiopia‘s existing laws and regulations should be assessed to
identify gaps and opportunities for integrating AI-specific regulations. The adaptation process
must ensure alignment with international standards while considering the country's unique
context. Building institutional capacity for AI governance will also be crucial, and Ethiopia
may need to invest in training and resources to effectively implement and enforce new
regulations.

Generally, by adopting and adapting international best practices to Ethiopia's unique context,
the country can create a strong AI governance framework. This approach will promote
innovation
while addressing ethical and regulatory challenges, helping Ethiopia navigate the
complexities of AI technology and harness its benefits while mitigating potential risk.