ANNOTATED BIBLIOGRAPHY 1-8 1
Course Project: Annotated Bibliography Sources 1-8
Taylor Ruiz
School of Education, Liberty University
ANNOTATED BIBLIOGRAPHY 1-8 2
Topic: The Ethical Impacts of Using AI-Driven Decision-Making in Education
Akgun, S., & Greenhow, C. (2021). Artificial intelligence in education: Addressing ethical
challenges in K-12 settings.7AI and Ethics,72(3).
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC8455229/
Five applications of AI are examined in this article along with their potential benefits for
K-12 education. As ethical challenges in K-12 education systems, privacy violations,
surveillance, autonomy issues, bias, and discrimination are discussed throughout the
article. Among the risks are data collection, tracking, predictive analysis, and automated
assessment algorithms, which perpetuate existing biases. This article examines the ethical
and societal risks associated with the integration of artificial intelligence into education.
Additionally, the author discusses various strategies and resources aimed at assisting
teachers and students in confronting ethical challenges, particularly those related to
privacy concerns and biases. Open-access resources on AI and ethics are available from
several research groups and nonprofit organizations, including MIT's "AI and Ethics"
curriculum and "AI and Data Privacy" workshop. Enhanced professional development
opportunities for K-12 teachers, as well as further research in the fields of teacher
education and educational technology, are emphasized by the author. Because of this
article's exposition of the ethical implications of artificial intelligence, and the use of this
technology in education, this article will be useful to my paper.
Baker, R. S., & Hawn, A. (2021). Algorithmic bias in education.7International Journal of
Artificial Intelligence in Education,732. https://doi.org/10.1007/s40593-021-00285-9
This article provides existing evidence, reviews theoretical perspectives, and provides a
framework for moving from unknown bias to known bias, this article reviews algorithmic
ANNOTATED BIBLIOGRAPHY 1-8 3
bias in education.7An analysis of algorithmic biases is provided by the author, as well as
the characteristics of the groups most affected.7Artificial intelligence bias discriminates
against many of the same protected classes that are protected by federal law.7There is
evidence that algorithms can be biased towards certain racial/ethnic groups, such as
African-Americans and Latinos.7These studies have largely excluded indigenous learners,
indicating a bias towards underrepresented contexts.7As far as bias mitigation is
concerned, judicial standards and legal frameworks can be used to assess the risks, while
past discriminatory patterns can be used to focus efforts on minimizing them.77AI systems
may not adequately recognize gestures of people with mobility or posture differences,
atypical speech patterns, or dyslexic spelling patterns because of algorithmic bias, which
has increased concerns about how it negatively impacts individuals with a range of
disabilities.7It has been found that models trained on one group do not perform well when
applied to a different group of learners.7The performance of learners can be improved by
collecting a diverse sample and training them, but there are still needs to be determined a
minimum size for the data sets.Researchers need to explore algorithmic bias in other
groups and determine which categories matter in various contexts. The author
recommends that learners' demographic data be collected, biases in labeling be avoided,
and distilled variables be used to address biases in training7data
to7improve7algorithmic7fairness in education.7The article will help me to write my paper
because it explains what algorithmic bias does to different groups of people.
Chen, Z. (2023). Ethics and discrimination in artificial intelligence-enabled recruitment
Practices.7Humanities and Social Sciences Communications,710(1), 1–12.
https://doi.org/10.1057/s41599-023-02079-x
ANNOTATED BIBLIOGRAPHY 1-8 4
This study examines how artificial intelligence is used in recruitment, how discrimination
can occur, and how initiatives have been taken to eliminate it. In addition to reviewing
literature, surveying respondents, and suggesting future directions, it offers suggestions
for further research. To assess current research on artificial intelligence algorithms for
recruitment, the author conducted a literature review. Several themes were identified in
the literature review, including AI-based recruitment applications, the causes and effects
of algorithmic discrimination, and the solutions to algorithmic discrimination in
recruitment. AI and algorithms are used to create automated systems, such as search
engines and job applicant screening systems, that analyze data to make predictions. As a
result, AI may duplicate prejudices in its decision-making based on these algorithms,
which contain biases. Due to biased datasets, engineer's biases, and feature selection,
algorithms can introduce bias into recruitment processes. Thus, high-tech systems
obscure the fact that they are not objective, perpetuating existing social prejudices and
discrimination. Machine learning and natural language processing techniques can be
discriminatory due to gender, race, skin color, and personality bias. One such example is
Amazon's machine learning-based hiring tool, Microsoft's chatbot Tay, and Google's
photo application algorithm. During the study, ten individuals who were experienced
with using AI-driven decision-making tools to interview or recruit were interviewed.
Live/video/telephone interviews lasted approximately 30 minutes each and were
conducted using face-to-face/video methods. In the interview process, six core questions
were asked about AI-driven recruitment, discrimination, causes, types, strategies, and
suggestions. Confidentiality, objectivity, and right-to-know principles were followed
during the interview process. A coding method was used to organize the data to derive
ANNOTATED BIBLIOGRAPHY 1-8 5
categories and nodes related to AI-driven recruitment and discrimination. According to
the author's analysis of the interviews, job seekers' perspectives on AI-driven hiring
applications, benefits for job seekers, concerns about discrimination, and
recommendations for preventing discrimination were revealed. The study is beneficial to
my paper since there are ethical concerns relating to hiring teachers and administrators
through AI-driven hiring applications. It can be detrimental to the diversity of schools if
these AI-driven hiring applications are used to select teachers, and it can lead to some of
the best teachers being passed over for a job due to algorithmic biases.
Greenstein, S. (2021). Preserving the rule of law in the era of artificial intelligence
(AI).7Artificial Intelligence and Law,730. https://doi.org/10.1007/s10506-021-09294-4
In this article, the author examines the contradiction between law and information
technology and the threat AI poses to the rule of law. An overview of the rule of law, an
overview of AI, and an analysis of how AI is impacting the rule of law are provided. AI
is an academic discipline that involves building intelligent artifacts. A wide range of
subjects are covered, including philosophy, mathematics, neuroscience, and linguistics.
Technology based on artificial intelligence is characterized by autonomy and
adaptability. Due to data availability, increased processing power, and advanced
mathematical algorithms, AI has become more popular in recent years. Machine learning
is a subcategory of artificial intelligence that uses algorithms to make autonomous
decisions using training data. The Loomis case highlights how technology can threaten
the rule of law. Among the five reasons the Supreme Court provided for caution are
proprietary software and reliance on national data. It is important to consider these factors
when developing artificial intelligence systems, as the author urges. The GDPR's
ANNOTATED BIBLIOGRAPHY 1-8 6
attempts to balance intellectual property rights against privacy rights in relation to AI
illustrate how technology can disrupt traditional law's balancing of interests. To remedy
this imbalance, various mechanisms have been proposed, such as trusted third parties.
Laws allow individuals to flourish, but artificial intelligence can subvert human agency.
The author explains that AI should be used to foster self-determination and social
cohesion, not undermine human flourishing. Using this article as an example, I am able to
explain that laws are designed for humans, which AI does not belong to. Furthermore, it
discusses the ethical implications of algorithmic bias on national data sets in AI
technology.
Holmes, W., Hui, Z., Miao, F., & Ronghuai, H. (2021).7AI and Education. UNESCO Publishing.
https://unesdoc.unesco.org/ark:/48223/pf0000376709
As AI and education become increasingly connected, this publication provides guidance
for policymakers on leveraging the opportunities and addressing the risks, including how
to ensure ethical, inclusive, and equitable use of AI in education. There is a huge impact
that artificial intelligence (AI) is having on the education sector. Education professionals
and policymakers need to work together to ensure that AI benefits all students, empowers
teachers, strengthens learning management systems, and prepares citizens to live and
work safely in an AI-dominated world. A deep learning-specific discussion of the
capabilities and limitations of current AI technology is presented in this publication.
Furthermore, this publication explains the numerous ways AI-driven decision making is
used across industries, including education, in addition to providing an in-depth analysis
of its current applications and ethical implications. As a result of this publication, I am
able to provide suggestions for solutions to some of the current ethical concerns related to
ANNOTATED BIBLIOGRAPHY 1-8 7
the use of AI in education along with a thorough analysis of the risks of AI in education
in my paper.
Khosravi, H., Shum, S. B., Chen, G., Conati, C., Tsai, Y.-S., Kay, J., Knight, S., Martinez-
Maldonado, R., Sadiq, S., & Gašević, D. (2022). Explainable artificial intelligence in
education.7Computers and Education: Artificial Intelligence,73, 100074.
https://doi.org/10.1016/j.caeai.2022.100074
This article introduces a framework for explainable AI in education, illustrated by four
case studies. A misuse of AI models may result in incorrect, incomplete, or
misinterpreted explanations, as well as dysfunctional behavior. For this reason, the author
urges designers to create simple and interpretable models, develop sound and complete
explanations, and incorporate best practices from game theory mechanism design. Also,
the article identifies research needs and opportunities to advance explainable artificial
intelligence in the classroom. A number of AI systems in education are studied using
explainable AI in the author's case studies. The case studies include examples of a system
for adaptive learning, a tool for analyzing writing, and a system for providing educational
healthcare. Using current research, the author argues that users should understand the
inner workings of AI systems in order to create trustworthy AI-augmented sociotechnical
systems based on outlined criteria. The article is beneficial to my paper because it
provides a framework for implementing artificial intelligence in education while also
showing the need for transparency about the inner workings of AI systems to build trust
with the end users.
ANNOTATED BIBLIOGRAPHY 1-8 8
Marcinkowski, F., Kieslich, K., Starke, C., & Lünich, M. (2020). Implications of AI (un-
fairness) in higher education admissions.7Proceedings of the 2020 Conference on
Fairness, Accountability, and Transparency.
https://doi.org/10.1145/3351095.3372867
In this study, algorithmic decision-making (ADM) versus human decision-making
(HDM) was compared in higher education admissions. Within the context of higher
education, this study examined students' perceptions of fairness regarding algorithmic
versus human decision-making. According to the results, algorithmic decision-making
was considered to be more fair, both in terms of procedural fairness and distributive
fairness, than human decision-making. Despite this, fairness concerns are gaining traction
in academic and public discourse since ADM systems may not be capable of making
objective, fair decisions. There are two dimensions of fairness when judging AI-driven
processes: procedural fairness and distributive fairness. In procedural fairness, the
process is evaluated, while in distributive fairness, the resources are considered.
Procedural fairness is judged in terms of consistency, neutrality, precision, revocability,
ethics, and representativeness, while distributive fairness is judged in terms of
discriminatory impact, reliability, relevance, and outcome. The study included 304
university students who answered questions about their attitudes toward and knowledge
of digitalization, rated the fairness of automated decision-making systems, and answered
questions about the reputation of their universities. Results showed that the perception of
fairness, both distributive and procedural, has an impact on students' intention to protest
ADM use and their likelihood of leaving. Furthermore, the study found that a university's
reputation is influenced by its perceived fairness in the ADM system. The study
ANNOTATED BIBLIOGRAPHY 1-8 9
contributes to my paper by showing students' views regarding the use of ADM in higher
education and demonstrating the need for a framework and regulation to prevent unfair
decisions being made in education.
Nasir, S., Khan, R. A., & Bai, S. (2023, August 31).7Ethical Framework for Harnessing the
Power of AI in Healthcare and Beyond. ArXiv.org.
https://doi.org/10.48550/arXiv.2309.00064
Deep learning methods and AI-driven decision-making have become increasingly
popular in real-world applications and industries over the past decade. With a particular
focus on health care, this article explores the ethical dimensions intricately linked with
the rapid development of AI technologies. According to the author, effective governance
and collaboration mechanisms are vital to the advancement of AI and its integration into
society. A six-pillar ethical framework for safe AI ecosystems is presented by the author,
promoting inclusivity, fairness, transparency, accountability, sustainability, resilience,
and security. For addressing the unique challenges of developing AI systems for
governance and collaboration, a comprehensive framework (Figure 9) is required. This
framework includes components such as Human Oversight and Intervention (HOI),
Multi-Stakeholder Engagement (MSE), Privacy by Design (PbD), Safe AI, Ethical
Governance Council (EGC), Awareness Programs, Continuous Improvement and
Adaptability (CIA), and Inter-Generational Considerations (IGC). HOI and PbD work
together to mitigate risks and ensure privacy in AI systems, while MSE promotes
inclusivity and interdisciplinary engagement. The EGC defines ethical principles and
influences awareness programs for safe AI that minimizes risks and unintended
outcomes. The CIA acknowledges the dynamic nature of AI governance, and the IGC
ANNOTATED BIBLIOGRAPHY 1-8 10
emphasizes the long-term effects of AI. As a result, ethical, transparent, and adaptable AI
systems are created that align with societal values and needs. This article makes a
significant contribution to my paper because it offers suggestions for mitigating ethical
risks associated with implementing AI-based technology while also explaining the need
for human intervention.
