Teaching Case

Journal of Information Technology

Teaching Cases
AI for learning unleashed: Pioneering 2024, Vol. 0(0) 1–10
© Association for Information
generative AI in education at the University Technology Trust 2024
Article reuse guidelines:
sagepub.com/journals-permissions
of Miami DOI: 10.1177/20438869241266258
journals.sagepub.com/home/jittc

Robert Wayne Gregory1  and Simran Narang1

Abstract
The University of Miami (UM) is at the forefront of integrating generative AI into education, spearheaded by Ann
M. Olazábal, Interim Dean of the Miami Herbert Business School. This case study explores the dynamic landscape of AI
adoption at UM, highlighting both the innovative potential and ethical challenges. The university has embraced AI tools like
Adobe Firefly and Copilot, aiming to enhance learning experiences and internal processes. However, the misuse of AI for
academic dishonesty poses significant dilemmas. The case contrasts two MBA students’ approaches to AI: Alex, who uses
AI to shortcut his studies, and Jordan, who leverages AI to deepen her understanding and enhance her learning. The
narrative underscores the need for ethical AI use and the importance of fostering critical thinking and creativity. Ann’s
strategic vision includes forming an AI task force to explore AI’s organizational impacts and potential use cases, such as a
smart assistant for student curriculum planning and an AI-powered entrepreneurship lab. The case concludes with a
discussion on balancing innovation with academic integrity and the challenges of centralized IT governance. The case serves
as a foundation for classroom discussion including a comprehensive examination of the opportunities and risks associated
with generative AI in higher education.

Keywords
Generative AI, artificial intelligence and neuroscience, innovation, business intelligence and analytics, new technology,
teaching cases

On a bright and sunny day in April 2024, the lush campus of learning and development. Ann’s commitment to ethical
the University of Miami (UM) in Coral Gables is abuzz with stewardship is unwavering, but she also acknowledges the
the usual academic fervor. In the heart of this scholarly transformative potential of AI in reshaping educational
oasis, Ann M. Olazábal, Interim Dean of the Miami Herbert experiences and processes within the school:
Business School, prepares to address the faculty and staff.
The gathering is convened for an update on the search for a “We have been teaching machine learning for a while, and now
new dean, but Ann has a broader vision to share before she more recently, the widespread use of generative AI tools in the
transitions back to her faculty role. She is poised to present a classroom have forced faculty to find ways how to adjust their
set of strategic recommendations aimed at positioning the assignments versus just using AI for cheating. When I think of
business school at the forefront of a university-wide ini- AI, I see the urgency to transform learning experiences for our
tiative: pioneering generative AI in education. customers while at the same time, we need to get our backend
The status quo of generative AI at the University of processes ready to support these customer-facing innovations
Miami reflects a dynamic intersection of innovation and as well. It’s just a fact that many of our internal processes could
caution. The institution has embraced the technology, in-
tegrating AI tools like Adobe Firefly and Copilot into its
educational fabric and fostering AI fluency among its
1
community. Yet, as Ann surveys the landscape, she rec- University of Miami - Coral Gables Campus, USA
ognizes the dilemmas that accompany this digital revolu-
Corresponding author:
tion. The misuse of generative AI by students for plagiarism Robert Wayne Gregory, University of Miami - Coral Gables Campus,
and cognitive shortcuts has surfaced, challenging the core 1320 S Dixie Hwy, Coral Gables, FL 33124-6914, USA.
values of academic integrity and the mission of human Email: [email protected]
2 Journal of Information Technology Teaching Cases 0(0)

be more efficient and supportive of customer-facing innovation applications where tasks are clearly defined and do not
using latest AI technologies.” require the creation of new content, such as data analysis,
structured decision-making, and automation of routine
As Ann contemplates the conclusion of her tenure, she is tasks. In contrast, generative AI learns from large datasets to
determined to leave a legacy that harmonizes the promise of understand underlying patterns and generate new data that
generative AI with the principles of the business school. The mimics these patterns, often using unsupervised learning
innovative use cases proposed by AWS—ranging from a techniques. This makes generative AI applicable to creative
smart assistant for student curriculum planning to fields such as digital art, music production, content gen-
leveraging AI for data analysis in entrepreneurship, and eration, and innovative product design.
enhancing online learning experiences—present tantalizing Generative AI builds upon predictive AI models, also
opportunities for educational innovation. Yet, the business called foundation models, but goes beyond predictive AI.
school’s limited autonomy in IT decisions, due to the While predictive AI focuses on forecasting future events or
university’s centralization efforts, poses a significant hurdle. behaviors based on historical data, by identifying patterns to
Ann envisions the business school as an agile entity capable make predictions about future outcomes, generative AI
of rapid innovation in partnership with leading technology leverages predictive capacities of machine learning algo-
providers, but she is acutely aware of the need to navigate rithms and foundation models but focuses on generating
the complexities of bureaucracy, risk mitigation, and ethical new content or data based on learned data patterns. It is
considerations. creative in nature, producing outputs like text, images, and
In her strategic recommendations, Ann seeks to chart a sounds that did not previously exist. In terms of data uti-
course that not only prioritizes the most promising use cases lization and technologies involved, generative AI goes
for generative AI but also establishes a framework for beyond predictive AI in that it uses a variety of machine
evaluating their impact on learning outcomes. Her vision is learning models, including Generative Adversarial Net-
clear: to empower the business school to lead by example works (GANs) and transformer models, to generate new
within the University of Miami ecosystem, ensuring that data that is similar in style or content to its training set. This
generative AI serves as a catalyst for learning, creativity, highlights that generative AI and predictive AI are not
and ethical growth. As she addresses her colleagues, Ann mutually exclusive and complement each other in complex
stands at the helm of change, ready to steer the business systems. For example, generative AI can create training data
school towards a future where technology and human po- for predictive AI models, enhancing their accuracy and
tential converge in harmony. robustness.

Generative AI A day in the life of two students

Generative AI and other forms of artificial intelligence (AI), In the bustling corridors of Miami Herbert Business School,
such as traditional symbolic systems and predictive AI, two students, Alex and Jordan, navigate their academic
differ primarily in their objectives, functionalities, and responsibilities with the aid of generative AI, albeit in
underlying technologies. starkly different manners. Their stories unfold as a testament
Gartner, Inc. defines generative AI in the following way: to the divergent paths technology can carve in the pursuit of
education.
“Generative AI refers to AI techniques that learn a represen-
tation of artifacts from data, and use it to generate brand-new,
unique artifacts that resemble but don’t repeat the original data.
Alex: The shortcut seeker
These artifacts can serve benign or nefarious purposes. Gen- Alex, a second-year MBA student, is counting down the
erative AI can produce totally novel content (including text, days until he can leave the academic pressure behind.
images, video, audio, structures), computer code, synthetic Unlike his peers, who often appear to have high self-esteem
data, workflows and models of physical objects. Generative AI and confidence in their own abilities, Alex’s approach to his
also can be used in art, drug discovery or material design.”1 studies is markedly different. He is often seen lounging in
the back of the classroom, disengaged from discussions and
Generative AI uses data-driven learning to generate lectures. “Assignments? Oh, I guess I’ll get to them
outputs that are similar to, but not exact replicas of, the input eventually,” he says with a shrug, barely concealing his lack
data it has been trained on. This includes generating text, of interest and laziness. His secret? “ChatGPT, I guess. I
images, music, and even code. This is distinct from tradi- have a subscription of course. It’s like having someone else
tional symbolic AI systems such as expert systems which do the degree for me.” His demeanor suggests not an air of
focus on performing specific tasks based on predefined rules confidence but a detachment from the academic rigor and an
and patterns. These traditional intelligent systems excel in underlying struggle to find motivation.
Gregory and Narang 3

For a typical assignment involving text-based answers, tools thoughtfully. A prime example is Jordan, another
Alex’s routine involves barely looking at the assignment MBA candidate, who employs generative AI judiciously,
brief and just copy-pasting it right into ChatGPT, and then blending skepticism with strategic use. “AI is a powerful
copy-pasting the answers provided by the machine for his tool, but it’s not without its flaws,” she remarks, reviewing a
assignment. “Why would I waste my time on a tedious draft on her laptop. Often, students find that relying on
assignment when AI can just do it for me?,” he muses, generative AI can lead to tunnel vision or confirmation bias,
genuinely impressed by the tool’s capabilities to predict and limiting their ability to generate original ideas. “I treat it as a
provide useful answers to the assignment prompts given by springboard to develop my thoughts and gain new insights,”
the professor. “We get so many assignments and have to she explains.
complete so many assessments, while I’d rather just get Jordan’s day is punctuated by sessions of prompt en-
school over with. Then, I’ll just learn anything I need to gineering, where she meticulously refines her questions to
know on the job once I start making money.” However, the AI, seeking not just any answer but the right one. “It’s
Alex did not realize how misusing AI could impact his about understanding the nuances, the context, and more
ability to learn and think critically. importantly making sure AI understands it too,” she ex-
His approach, however, is not without its critics. “Alex plains. “I cross-check facts and challenge assumptions
relies too heavily on AI,” remarks a classmate, who prefers based on the knowledge I have acquired in my courses and
to remain anonymous. “He’s missing out on learning, and through additional self-study and reading. That’s how you
not only that, he’s affecting our learning and grades on grow, not by taking cognitive short-cuts and focusing on
group projects too!” But Alex remains unfazed, confident in getting things done overly quickly, but instead taking the
his method. “As long as I can pass, I don’t care if I get time to immerse yourself in the course content and as-
reprimanded” he asserts, his trust in the AI’s output signments to learn every day.”
unwavering. Her peers admire her diligence. “Jordan doesn’t take
University of Miami’s professors have to deal with shortcuts,” says a study group member. “She uses AI as a
students like Alex. They are increasingly troubled by the tool, not a crutch. It seems to make her more efficient at
rise of AI-assisted cheating. Robert Plant, Associate Pro- making valuable contributions to team assignments without
fessor of Business Technology in the Miami Herbert affecting the authenticity of her work.” This sentiment is
Business School, complains about students like Alex: “We echoed by her professors, who have noticed her deep en-
have to uphold the honor code at the university. Cheating is gagement with the material. “Her work reflects a genuine
not allowed, period. I can’t tolerate students who cheat on desire to learn,” one professor notes. “She is leveraging
their assignments by using these fancy new tools.” technology to enhance her understanding and improve her
The challenge, however, is that the professors struggle to capabilities, not bypass them.”
prove when AI has been used, as the technology’s outputs As the sun sets on the Coral Gables campus of the
can be indistinguishable from student-generated work. This University of Miami, the stories of Alex and Jordan
inability to confirm suspicions hampers their efforts to highlight a critical juncture in modern education. One
maintain academic integrity. Professors are thus caught in a student seeks convenience above all, while the other pur-
dilemma between enforcing standards and the limitations of sues knowledge with technology as her guide. Both are
detection methods. They advocate for rethinking assess- charting their courses in a world where AI’s role in learning
ment strategies to emphasize creativity and critical thinking, is still being defined, their narratives a mirror to the choices
which are harder for AI to replicate. that lie ahead for all students in the digital age.
Ola Henfridsson, Professor of Business Technology and
a colleague of Robert Plant, thinks that essays are no longer
an option: “If I see my MBA students submitting entire
Generative AI @University of Miami
essay-style assignments which look amazing at first sight Shifting from the student to the faculty perspective on
but are full of inaccuracies and inconsistencies with things generative AI, Ann felt that it was time to act and put to-
we actually discussed in the course, I get the feeling we have gether an AI task force in the business school to discuss the
to come up with new assessment methods.” Additionally, organizational impacts of AI and how to best leverage the
there is a call for clearer guidelines on acceptable AI use in new technology while mitigating the risks involved and
academic settings, aiming to balance technological benefits following the defined principles (see Exhibit 1). As faculty
with ethical academic practices. Robert W. Gregory, Associate Professor of Business
Technology, approached Ann with questions regarding the
design of new executive education offerings focused on AI,
Jordan: The thoughtful collaborator she reflected: “I really think we cannot be complacent. I
On the other hand, both Robert and Ola observed many know we are doing well in the business school, and the
instances in their classes of students using generative AI university overall, but I sense some degree of urgency to
4 Journal of Information Technology Teaching Cases 0(0)

embrace AI more seriously to outperform our competition rules we have for study and curriculum planning as well as
and consolidate our leading market position.” degrees formulated in a way that they are codifiable using
These comments echoed public statements made by the current technology.”
David Chapman, a computer science professor at UM Robert P.’s view in the meeting was that the rules
emphasizing that people who refuse to become fluent with knowledge could be made more easily accessible and ap-
AI in day-to-day use could suffer. “You risk being left plicable for each student: “If you are a faculty parent, we
behind,” he said. “From now on, it’s going to be part of our know how to expedite the process because we know the
lives and it can make people more effective at what they rules, or at least know who to ask and find out how to apply
do.”2 Ann convened the first meeting of the strategic AI task them. This can help expedite the process to get to gradu-
force on May 1st, 2024. ation quicker, and I got my son all the way through to his
Allan Shawn Gyorke, Assoc. VP, Information Tech- degree in three years instead of four. But these rules are not
nology, Asst. Provost for Educational Innovation, and Chief equally accessible to everyone right now, at least not
Academic Technology Officer at UM, gave an overview of through an intuitive and easy-to-use interface like
what the university already does with AI as of May 2024. ChatGPT.”
He explained: “UM is actively engaged in exploring the use Robert G. intervened in the meeting with a question why
of generative AI across various domains. The strategic so much focus thus far had been on adopting off-the-shelf
vision is one of encouragement for innovation, research, AI software rather than building UM-specific AI software
and education.” For instance, the Frost Institute for Data trained on proprietary data to achieve more meaningful
Science and Computing (IDSC) at UM announced a outcomes with AI. This quickly triggered a discussion about
funding opportunity specifically for generative AI research, the risks involved in adopting AI in such a more advanced
envisioning awards to support projects that explore the use way. One key concern was cybersecurity, which the cen-
of generative AI in ethical, educational, scientific, business, tralized IT department at UM prioritized recently due to
medical, engineering, law, and humanities contexts.3 In ongoing threats and concerns about potential attacks. Ste-
general, the institute offers support for faculty working on phen Scheel, Executive Director of Information Technol-
grant-related research projects to train and develop machine ogy, explained the tension in the following way: “We are at
learning models and other AI applications needed as part of the cusp of all other universities starting to leverage AI in
the research project. much more innovative ways, while the situation here is that
According to Allan, the UM Information Technology the pendulum has swung all the way from being innovative
department (UMIT) provides access to a range of different in the other direction of focusing on cybersecurity and
AI tools4: (1) Adobe Firefly, an AI-powered tool focused on mitigating the risks of new technologies driven by the
enhancing creativity by providing new ways to imagine, consumer market, and until we get a good handle of these
experiment, and bring ideas to life; (2) Copilot, another AI- threats I don’t think we will move much.”
powered tool focused on getting work done faster, being The first AI task force meeting ended with a brief dis-
more creative, or supporting customers better; (3) Grade- cussion about potential AI use cases developed in part-
scope, a tool offering AI-assisted grading support; (4) nership with AWS (see next section). It was not clear to the
Blackboard AI Design Assistant, a tool that helps educators task force members, however, how to evaluate these dif-
co-create essential course elements like module titles, de- ferent ideas and which ones to prioritize. A key question
scriptions, images, test questions, prompts for assignments, was also how AI technology could be used intelligently to
and rubrics for grading. support engaged students like Jordan and educate students
Even though members of the AI task force acknowl- to avoid the path taken by Alex.
edged the advancements made in giving faculty, staff, and
students access to off-the-shelf AI tools, tensions surfaced in
the meeting. One concern expressed was that many internal Partnership with AWS: Three potential
processes at the university level as well as within the use cases
business school itself still relied heavily on manual steps
U-Pilot: Your AI academic advisor
and were not always consistent and reliable. For example,
Ann pointed out in the meeting that progress reporting and A potential use case for adopting generative AI in part-
degree audits for students were cumbersome and not always nership with AWS (see Exhibit 2) is to develop a smart
predictable: “It is really difficult other than doing it by hand assistant for students at the University of Miami to help with
that a student has everything they need. The rules for what study and career planning: The University of Miami could
courses you can choose, and in which sequence, while most leverage AWS’s generative AI services like Amazon
of them are in the bulletin, not all of them, they are certainly Bedrock5 (see Exhibit 2) to build a conversational AI as-
complex and not all built into the programmed software sistant that provides personalized study advice and career
tools we currently use.” Allan added: “Not always are the guidance to students. This assistant would utilize large
Gregory and Narang 5

language models and natural language processing to engage learning resources tailored to the student’s courses, rec-
in human-like dialog with students. ommends tutoring and study groups for challenging sub-
Some key capabilities and innovations could include the jects, and sends reminders for upcoming tests and project
following: (1) providing tailored study tips and learning deadlines. The career exploration tool assesses the student’s
strategies based on the student’s specific courses, learning career interests, skills, and values through quizzes and
style, strengths, and weaknesses.6 The AI could analyze the conversations; suggests potential career paths and job roles;
student’s academic record and course syllabi to offer tar- and provides links to information on qualifications, salary
geted advice; (2) helping students explore career options data, and employment outlooks. To aid in skill building, the
aligned with their interests and skills. The assistant could platform recommends extracurricular activities, research
ask questions to understand the student’s goals and aptitudes projects, and internships that enhance professional skills;
and then suggest relevant career paths and job roles to assists in applying to these opportunities; and tracks skill
consider; (3) recommending specific actions to build skills development over time. Additionally, it connects students
and gain experience, such as research projects to pursue with alumni mentors in their fields of interest for infor-
with faculty, student organizations to join, or internships to mational interviews and facilitates mentor relationships for
apply for based on the student’s profile and aspirations; (4) career insights and advice. Finally, the job search func-
answering common questions about university policies, tionality offers tools to create tailored resumes and cover
course registration, financial aid, and more—serving as an letters, suggests job and internship openings that match the
always-available advisor to guide students through ad- student’s qualifications, and provides interview tips and
ministrative processes; (5) analyzing job market data and practice questions.
alumni employment outcomes to provide data-driven in-
sights into career prospects for different majors and path-
Generative AI entrepreneurship lab
ways the student may be considering.
To build this, a cross-functional University of Miami UM could create a Generative AI Entrepreneurship Lab
team including faculty from computer science, education, housed within its Miami Herbert Business School. The lab’s
and career services could partner closely with the AWS ML focus would be to immerse students in using generative AI
team and solutions architects. They would use AWS AI/ML to solve real-world business problems. Professors teaching
tools and infrastructure to develop, train, and deploy the courses at the intersection between business technology,
models powering the assistant. Key steps would include business functions, and entrepreneurship would leverage
gathering training data from course content, academic the lab’s resources and partnerships with technology firms
records, and career resources; defining the knowledge base and startups to provide students hands-on experience
and dialog flows; testing and validating the models; and building and deploying generative AI models. A strategic
integrating the assistant into student-facing apps and por- partnership with AWS (see Exhibit 2) would provide the
tals. AWS’s pre-trained models and development acceler- foundational technological capabilities for the lab. AWS
ators could enable a rapid implementation. infrastructure could be used to analyze large datasets and
This initiative would be a pioneering example of using generate solutions to business problems, enhancing stu-
generative AI to provide scalable, personalized academic, dents’ experiential learning outcomes.
and career support to students, tailored to their individual Additionally, the lab could integrate Amazon Q De-
context and goals. It could enhance the student experience, veloper, an AI-powered coding companion, to help students
boost retention and employment outcomes, and establish the learn and improve their coding skills. The technology
University of Miami as a leader in AI-powered education provides real-time suggestions, code completions, and
innovation. explanations as students write code, making it an invaluable
The app, called “U-Pilot: Your AI Academic Advisor,” tool for those new to programming or looking to enhance
would have a clean, modern interface optimized for mobile their coding abilities.
devices7 (see Exhibit 3). The platform features a simple The learning experience using the lab could be structured
chat interface where students can interact with an AI as- as follows: (1) The professor defines a specific business
sistant. This assistant offers personalized advice and in- problem or use case, such as customer analytics, person-
formation, citing relevant sources, and includes buttons for alized marketing, fraud detection, or risk management. (2)
easy access to key features such as course planning and Students work in teams to gather and feed large amounts of
career exploration. In the academic planning section, the relevant business data into a generative AI application
system displays the student’s course schedule and degree hosted on AWS. This could leverage AWS services like
progress, recommends optimal courses based on the stu- Amazon S3 for data storage, Amazon SageMaker for model
dent’s major, interests, and goals, and assists in mapping out training and deployment, and Amazon EC2 GPU instances
future semester plans to ensure timely graduation. For study for high-performance computing. (3) Students use gener-
support, it provides customized study strategies and ative AI techniques like data augmentation, prompt
6 Journal of Information Technology Teaching Cases 0(0)

engineering, and fine-tuning foundation models to train Learning Pathways: Generative AI models analyze stu-
their AI application to analyze patterns in the data and dents’ skills, experience, career goals, and progress to
generate high-quality solutions to the defined business recommend customized learning modules, resources, and
problem. (4) As students develop the code for their AI activities. The platform continuously adapts content and
applications, they can utilize Amazon Q Developer to assist pacing to optimize each student’s mastery. (3) AI Teaching
them in writing efficient and accurate code. Q Developer’s Assistants: Students get 24/7 access to AI chatbots that can
suggestions and explanations can help students learn best answer questions, provide explanations, and offer guidance.
practices and understand the logic behind their code. (5) The AI learns from all student interactions to give in-
Throughout the process, students gain practical experience creasingly relevant support. (4) Immersive Case Simula-
with AWS infrastructure, data preprocessing, model eval- tions: Leveraging AWS’ metaverse and VR/AR services,
uation, and addressing ethical considerations of AI systems. the platform generates realistic business case scenarios
(6) The lab would leverage existing space in the entre- where students collaboratively solve problems, applying
preneurship area (e.g., the “Fish Bowl”), allowing student concepts from coursework. Generative AI creates dynamic
teams to pitch their generative AI solutions to the business characters and challenges that adapt to student actions. (5)
problem, explaining their approach and results, similar to an Personalized Career Coach: Generative AI provides each
entrepreneurial venture pitch. student with an AI career advisor that offers job recom-
The Generative AI Entrepreneurship Lab would give mendations, resume feedback, interview prep, and net-
students valuable experience applying cutting-edge AI working tips tailored to their profile and goals. (6) Adaptive
technology to real business cases using enterprise-grade Assessments: AI-generated quizzes and assignments can
AWS tools while also improving their coding skills with automatically adjust question types and difficulty to match
Amazon Q Developer.8 It builds their skills in data analysis, each student’s knowledge level, delivering a customized
solution design, model development, AWS, coding, and challenge. Instant grading and targeted feedback highlight
entrepreneurial problem-solving. Students could also le- areas for improvement, motivating students to revisit ma-
verage the acquired skills in applying for jobs at the in- terial and try again.
tersection of technology or within the technology sector Under the hood, the platform would be powered by
itself. Hosting the lab application and computing on AWS AWS’s comprehensive generative AI services and infra-
provide students access to the necessary AI infrastructure structure: (1) Amazon Bedrock to access and fine-tune
while learning cloud and data skills that are in-demand in foundation models on Miami Herbert’s proprietary data
the job market. and content, (2) Amazon SageMaker to build and deploy
A key benefit of the lab would be to prepare students for custom ML/DL models, (3) Amazon EC2 for scalable
the business applications of AI while seeding the next compute, and (4) Amazon S3 for secure data storage.
generation of AI entrepreneurs. It’s an innovative way to Amazon SageMaker is a fully managed machine learning
integrate the latest technologies into the curriculum so platform that enables developers and data scientists to
students gain the technical skills, coding abilities, and quickly and easily build, train, and deploy ML models at
business acumen to thrive in an AI-driven world. In building any scale. In the context of the Miami Herbert AI-Powered
the lab and offering hands-on courses that leverage the lab’s MBA Learning Platform, SageMaker would be used for
capabilities, UM would follow other universities that are several key functions, including data preparation, model
already doing this9 based on best practices of generative AI training, model hosting, model monitoring, and automated
on AWS.10 ML workflows. As Miami Herbert Business School cur-
rently partners with 2U which has outdated technology, the
new partnership with AWS would mean phasing out the
Miami Herbert AI-Powered MBA learning platform existing partnership and not renewing any existing con-
The Miami Herbert Business School could partner with tracts. Infusing generative AI across the learning journey
AWS (see Exhibit 2) to build a cutting-edge online learning will allow the school to deliver a premium, highly differ-
platform for its MBA program powered by generative AI. entiated online MBA experience that drives student en-
This platform would provide a highly engaging and per- gagement, outcomes, and loyalty.
sonalized learning experience tailored to each student’s
unique needs, interests, and learning style.
Key features could include the following: (1) Intelligent
Outlook and next steps
Course Designer: Professors use generative AI tools to As UM continues to integrate generative AI into its cur-
rapidly develop dynamic, interactive course content. The AI riculum, the institution stands at a critical juncture of in-
can generate lesson outlines, discussion questions, case novation and ethical responsibility. Under the interim
studies, and assessments based on learning objectives and leadership of Ann M. Olazábal at the Miami Herbert
source material provided by the professor. (2) Personalized Business School, the university is poised to lead a
Gregory and Narang 7

transformative initiative in generative AI education. Despite 2. The use of AI must be open and documented. The
the challenges posed by potential misuse of AI technologies use of any AI in the creation of your work must be
for academic dishonesty, the university recognizes the declared in your submission and explained. Your
profound opportunities AI presents for enhancing educa- faculty can provide guidance as to the format and
tional processes and outcomes. contents of the disclosure. The undeclared use of any
Ann’s strategic vision emphasizes the dual imperative of AI (including text, images, program code, and
leveraging AI to enrich learning while rigorously upholding musical notation) in any work may be considered as
academic integrity. She advocates for the development of AI plagiarism.
applications that not only improve administrative effi- 3. Engage with AI Responsibly and Ethically. Engage
ciencies but also directly enhance student learning expe- with AI technologies responsibly, critically evalu-
riences. This includes the use of AI in customizing ating AI-generated outputs and considering potential
curriculum planning, analyzing data in entrepreneurial biases, limitations, and ethical implications in your
studies, and refining online educational platforms. How- analysis and discussions. Ensure that the data used
ever, the centralized IT governance at the university poses for AI applications are obtained and shared re-
challenges to the rapid deployment of these innovations. sponsibly. Never pass off as your own work gen-
Some key questions remained. How could UM balance erated by AI.
the innovative potential of generative AI with the need to 4. You are 100% responsible for your final product. You
maintain academic integrity and prevent its misuse for are the user; if the AI tool makes a mistake, and you
academic dishonesty? In what ways could UM ensure that use it, then it’s your mistake. If you don’t know
the integration of generative AI into its curricula enhances whether a statement about any item in the output is
rather than diminishes the human learning experience, true, then it is your responsibility to research it. If you
fostering creativity and critical thinking skills among stu- cannot verify it as factual, you should delete it. You
dents? Considering the challenges of centralized IT gov- hold full responsibility for AI-generated content.
ernance, what strategies could the Miami Herbert Business Ideas must be attributed, and sources must be verified.
School use within the larger UM context to advocate for and 5. These principles are in effect unless the instructor
implement AI technologies that align with the business gives you specific guidelines for an assignment or
school’s specific educational goals and needs? exam. It is your responsibility to ensure you are
following the correct guidelines. Not following them
will result in a breach of the Academic Integrity
Preparation questions Policy.
6. Data that are confidential or personal should not be
Q1. What are the opportunities and challenges of gen-
entered into generative AI tools. Putting confidential
erative AI for the university and for students?
or personal data into these tools exposes you and
Q2. How does generative AI usage impact human
others to the loss of important information. There-
learning and development?
fore, do not do so. See point 3 above.
Q3. Which of the three generative AI use cases should
7. The rules and practices on the use of AI may vary
UM pursue? Why? What criteria would you use to
from class to class, discipline to discipline. Do not
assess and compare the use cases?
assume that what is acceptable in a Computer Sci-
Q4: What are your recommendations for the dean of
ence class will be acceptable in a Philosophy class. It
MHBS? How to move forward?
is the student’s responsibility to stay informed as to
the instructor’s expectations. When in doubt, ask.
Exhibits Exhibit 2: The Generative AI Landscape
Exhibit 1: University of Miami’s Principles for the use of
artificial intelligence (AI) tools.11 The generative AI landscape has continued to evolve
rapidly since the release of ChatGPT, with major technology
1. AI should help you think, not think for you. AI tools firms and a host of startups driving forward innovations and
may be used to help generate ideas, frame problems, expanding the applications of Large Language Models
and perform research. It can be a starting point for (LLMs) and other generative technologies. Major compa-
your own thought process, analysis, and discovery. nies continue to invest heavily in AI research and devel-
Do not use them to do your work for you, for ex- opment, driving forward the capabilities and applications of
ample, do not enter an assignment question into AI technologies. The competitive landscape includes a mix
ChatGPT and copy and paste the response as your of established tech giants and agile startups, each con-
answer. tributing to the evolution of generative AI.
8 Journal of Information Technology Teaching Cases 0(0)

Amazon Web Services (AWS) has developed a com- generative AI applications with security, privacy, and re-
prehensive suite of offerings related to generative AI, de- sponsible AI. Using Amazon Bedrock, you can easily exper-
signed to support a wide range of applications across iment with and evaluate top FMs for your use case, privately
various industries. These offerings are built on AWS’s customize them with your data using techniques such as fine-
robust cloud infrastructure, providing security, scalability, tuning and Retrieval Augmented Generation (RAG), and build
and accessibility. Here are the key components of AWS’s agents that execute tasks using your enterprise systems and data
generative AI capabilities: sources. Since Amazon Bedrock is serverless, you don’t have to
Foundation Models and Customization: AWS pro- manage any infrastructure, and you can securely integrate and
vides access to leading foundation models that can be deploy generative AI capabilities into your applications using
customized with customers’ proprietary data. This cus- the AWS services you are already familiar with.”12
tomization allows businesses to tailor AI models to their
specific needs, enhancing the relevance and effectiveness of Amazon SageMaker facilitates the building, training,
the AI applications. and deployment of machine learning models at scale. It
Amazon Q Developer assists developers and IT pro- includes SageMaker JumpStart, which helps users discover,
fessionals with all of their tasks—from coding, testing, and explore, and deploy open-source foundation models or
upgrading applications, to troubleshooting, performing create their own. This service provides managed infra-
security scanning and fixes, and optimizing AWS resources. structure and tools to accelerate the development of reliable
This service is available for free to individual developers, and secure AI models.
with paid tiers offering additional features like enterprise Infrastructure and Performance: AWS provides high-
security. performance GPU-based Amazon EC2 instances and
Amazon Q Business empowers companies’ employees purpose-built accelerators like AWS Trainium and AWS
to be more data-driven and make better decisions using Inferentia. These resources are optimized for training and
company data. It is a generative AI-powered assistant that running large machine learning models, offering cost-
can answer questions, provide summaries, generate content, effective and energy-efficient solutions for generative AI
and securely complete tasks based on data and information workloads.
in enterprise systems. It empowers employees to be more Other Players: OpenAI, initially established as a non-
creative, data-driven, efficient, prepared, and productive. profit, has continued to lead in the generative AI space with
Amazon Q Business also brings its advanced generative AI its advanced models. The company offers GPT-4, which has
technology to Amazon QuickSight, AWS´s unified Busi- been integrated into various applications including
ness Intelligence (BI) service built for the cloud. It allows ChatGPT and DALL-E 3. OpenAI’s models are known for
business analysts to use natural language to build BI their ability to generate human-like text and creative con-
dashboards in minutes and easily create visualizations and tent. The company has also introduced Sora, a text-to-video
complex calculations. tool, expanding its capabilities into multimedia content
Amazon Q Apps enables companies’ employees to use generation.
natural language to securely build their own generative AI OpenAI has significantly deepened its commercial
applications from their companies data, without any coding partnerships, particularly with Microsoft. Following the
experience. Employees simply describe the type of app they integration of GPT-3.5 and the subsequent GPT-4 into
want in natural language, and Q Apps will quickly generate Microsoft’s cloud services and products like Bing and
an app that accomplishes their desired task, helping them Microsoft 365, these collaborations have continued to
streamline and automate their daily work with ease and evolve. Microsoft’s investment in OpenAI has reportedly
efficiency. reached around $10 billion, emphasizing the strategic im-
Amazon Bedrock is a fully managed service that allows portance of this partnership for both entities.13 Microsoft
users to experiment with various top foundation models has leveraged OpenAI’s models to enhance its Azure cloud
(FMs), customize them with private data, and create platform, offering AI solutions that are increasingly em-
managed agents to execute complex business tasks. This bedded across its suite of products. This includes AI co-
service simplifies the process of building and scaling pilots in Microsoft Office applications and enhancements in
generative AI applications. According to Amazon’s its Bing search engine, which now integrates generative AI
website: to provide more context-rich search results.
Google has responded to competitive pressures by ac-
“Amazon Bedrock is a fully managed service that offers a celerating its own generative AI initiatives. Google’s AI
choice of high-performing foundation models (FMs) from division, including teams from Google Brain and Deep-
leading AI companies like AI21 Labs, Anthropic, Cohere, Mind, has been pivotal in driving AI innovation at Google.
Meta, Mistral AI, Stability AI, and Amazon through a single Google has developed and deployed several AI models,
API, along with a broad set of capabilities you need to build including Gemini and Vertex AI, a fully managed, unified
Gregory and Narang 9

AI development platform for building and using generative focuses on maximizing the accessibility of AI tech-
AI. Gemini for Google Workspace integrates AI into pro- nologies and has established a significant user base
ductivity tools. Google’s approach emphasizes enhancing with applications in imaging, language, and other
its existing suite of applications with AI to improve user domains.
experience and functionality. · Jasper AI provides AI-powered content generation
Google has played a pivotal role in the development and tools that assist with writing and creative tasks. It is
advancement of generative AI through several foundational designed to enhance productivity by automating
innovations, particularly with the introduction of the Trans- content creation, making it a valuable tool for mar-
former architecture. The Transformer architecture, introduced keters and content creators.
by Google in 2017 in the seminal paper “Attention is All You · Perplexity AI is a generative artificial intelligence
Need,” has been a cornerstone in the field of natural language search engine startup that has made significant strides
processing (NLP) and generative AI. This architecture is in the AI industry by offering a unique conversational
distinguished by its use of self-attention mechanisms, which AI platform. This platform allows users to interact
allow models to weigh the importance of different words in a with the search engine in a natural, human-like
sentence, regardless of their distance from each other in the manner, providing direct answers rather than just a
text. This capability enables the Transformer to capture list of links.
complex linguistic nuances and generate more coherent and · Ello is an innovative startup that has made significant
contextually relevant text outputs. strides in the education sector by leveraging gener-
NVIDIA remains a key player in providing the hardware ative AI to enhance childhood literacy. Ello has de-
and software necessary for AI development. Its offerings veloped a proprietary AI engine called Adaptive
include NVIDIA AI, NVIDIA NeMo, NVIDIA BioNeMo, Learn™, which functions as an advanced reading
and NVIDIA Picasso, along with various chips and GPUs coach for children in kindergarten through third
that are crucial for AI processing. grade. The platform is designed to listen, understand,
Meta has developed AI models like LLaMA and tools for and interact with children as they read, providing real-
generative AI applications across its social media and com- time support and feedback to improve their reading
munication platforms. The company focuses on integrating AI skills.
to enhance user interactions and content personalization.
Snowflake released Arctic, an open-source LLM de- Generative AI Startups as of 2023 (Source: CB Insights).
signed to cater specifically to enterprise-grade AI applica-
tions which require robust, scalable AI solutions.
Several startups are making significant contributions to
the generative AI field:

· Anthropic is known for its Claude platform, a large

language model that competes with OpenAI’s
ChatGPT. Claude is designed to be customizable and is
noted for generating less inappropriate content. An-
thropic’s approach includes a focus on AI safety and
ethics.
· Cohere specializes in building multilingual large lan-
guage models for enterprises, enhancing the capabilities
of chatbots, knowledge assistants, and search solutions.
Its recent funding rounds and rapid growth underscore
its potential impact on enterprise AI applications.
· Hugging Face is a collaborative AI community Exhibit 3: U-Pilot Wireframe (Generated with Uizard).
platform that provides tools for developers to build
and deploy AI models. It has become a central hub for
AI developers, offering over 61,000 pre-trained
models and 7,000 datasets. It supports a wide
range of AI applications through its collaborative
platform, contributing significantly to the democra-
tization of AI technology.
· Stability AI has made waves with its open-source
image generator, Stable Diffusion. The company
10 Journal of Information Technology Teaching Cases 0(0)

Declaration of conflicting interests 9. https://ep.jhu.edu/courses/635674-generative-ai-for-

entrepreneurs/https://www.heinz.cmu.edu/current-students/
The author(s) declared no potential conflicts of interest with re-
courses/94-844/
spect to the research, authorship, and/or publication of this article.
10. https://www.oreilly.com/library/view/generative-ai-on/
9781098159214/https://aws.amazon.com/developer/
Funding
generative-ai/
The author(s) received no financial support for the research, au- 11. https://petal.miami.edu/resources/using-ai-for-teaching-
thorship, and/or publication of this article. learning-and-scholarship/index.html
12. https://aws.amazon.com/bedrock/
ORCID iD 13. https://www.nytimes.com/2023/01/23/business/microsoft-
Robert Wayne Gregory  https://orcid.org/0000-0002-4843-5500 chatgpt-artificial-intelligence.html

Notes
Author biographies
1. https://www.gartner.com/en/information-technology/
Robert Wayne Gregory is an associate professor at Miami
glossary/generative-ai
Herbert Business School, University of Miami. He received
2. https://www.cbsnews.com/miami/news/artificial-intelligence-
his Ph.D. in management information systems from Goethe
could-wipe-out-education-gaps-between-students/
University Frankfurt. His research interests include digital
3. https://idsc.miami.edu/call-for-proposals-in-generative-ai/
innovation and strategy, platforms, and technology-driven
4. https://www.it.miami.edu/about-umit/resources/ai-tools/index.html
change.
5. https://aws.amazon.com/bedrock
6. https://aws.amazon.com/education/ed-tech/ Simran Narang is a master student in Management of
7. https://aws.amazon.com/blogs/mobile/connecting- Technology at the Engineering School, University of
applications-to-generative-ai-presents-new-challenges/ Miami. Her research interests focus on the intersection of
8. https://aws.amazon.com/q/developer/ artificial intelligence, learning, and business applications.