1 Introduction to Artificial Intelligence

in Higher Education

One small input for AI, one giant outcome for humanity.
Cecilia KY Chan

1.1 Introduction
The advancement of Artificial Intelligence (AI), Big Data, and the Internet of Things (IoT)
has disrupted many industries and, likewise, has the potential to revolutionise the field of
education. One tool that has been making waves in this regard is the Chat Generative Pre-
trained Transformer – best known as ChatGPT, a pretrained language processing model
developed by OpenAI. In this chapter, we will first introduce the concepts of AI, Big Data,
IoT and Generative AI (GenAI) to offer our readers a basic understanding of these tech-
nologies, providing some concrete examples from educational settings. To equip readers
with the necessary knowledge to make informed decisions on incorporating AI into the
classroom, this chapter will also go over some fundamental technical aspects of AI. We will
then delve into why ChatGPT is a game changer for higher education, what the tool is
capable of, and why it has suddenly gained such significant attention. This chapter will
conclude with questions that we may have regarding the application of AI and ChatGPT
in education.

1.2 Artificial Intelligence (AI)

When thinking about AI, we might immediately associate it with science fiction and shiny
robots; in fact, however, the science of AI has been around for decades. As per the term
Artificial Intelligence (AI) itself, it is a man-made system of computer-based algorithms
and programmes that can perform tasks typically requiring human intelligence. We can
find it in software programmes that we use in our everyday lives, from platforms like
YouTube, Amazon and Taobao, Meta (Facebook), Google Translate, Google Maps,
Microsoft Bing and more. These “intelligences” are built into such systems that have capa-
bilities in image processing, speech recognition, decision-making, language translation,
and even artificial creativity.
AI is the simulation of human intelligence in machines that are designed to think and
act like humans. AI involves the development of algorithms and models that enable com-
puters to learn from data and make predictions or decisions based on said data.
Advancements in electronics and computing hardware over recent years have allowed for
huge leaps in the development of AI, enabling faster computing power, increased data and
memory storage, and the creation of new algorithms such as deep learning.

DOI: 10.4324/9781003459026-1
This chapter has been made available under a CC-BY-NC-ND license.
2 Introduction to Artificial Intelligence in Higher Education
1.2.1 Three Types of AI

Artificial Intelligence has come a long way since its inception in the 1950s. According to
researchers and AI philosophers, AI has already gone through significant change and
expansion (e.g., Cordeschi, 2007; Shao et al., 2022), with new advancements continually
building upon prior developments to produce increasingly sophisticated systems. Currently,
discourse has identified three broad categories of AI, based on the complexity and extent
of machine intelligence: Artificial Narrow Intelligence (ANI), Artificial General Intelligence
(AGI), and Artificial Super Intelligence (ASI). Figure 1.1 shows the types of AI. While
these can be considered as the evolutionary stages of AI, it should be noted that AGI and
ASI are still theoretical concepts and not necessarily achieved through linear progression.
Artificial Narrow Intelligence (ANI) refers to systems that are designed to perform spe-
cific tasks and are trained on large datasets to execute the tasks with precision, such as
playing chess or translating languages. However, they are not capable of performing tasks
that require human-level intelligence. ANI has helped us make numerous breakthroughs in
society, such as in the domains of speech recognition and object recognition or detection
(LeCun et al., 2015). Siri, Alexa, fraud detection systems which analyse financial data for
patterns and anomalies, as well as self-driving cars that use data from sensors, cameras and
GPS to control vehicles, are all examples of ANI. These systems are often referred to as

ANI AGI ASI

- Weak AI - Strong AI - Machine

- Machine Learning - Machine Intelligence Consciousness
- Specialises in one area - Designs to mimic - Surpasses human
and solves one problem human intelligence and intelligence in every
behaviour possible aspect
- Replicates human
behaviour based on a - Possesses the ability - Possesses the ability
limited set of factors to understand or learn to outsmart human
and situaons any intellectual task beings in all tasks,
that a human can including higher order
- Examples: Chess AI, thinking skills such as
Email Spam filters, - Example: An AI creavity, general
Photo Enhancer teacher knowledge, and
problem-solving ability

Figure 1.1 Three Types of Artificial Intelligence.

 Introduction to Artificial Intelligence in Higher Education 3
“weak AI” (Thierer et al., 2017) as they have limited capabilities and lack general ‘intelli-
gence’ due to being optimised for narrowly-defined tasks.
Artificial General Intelligence (AGI) refers to machine systems that will become capable
of performing any intellectual task that humans are capable of. AGI refers to a hypotheti-
cal future stage of AI which “[bridges] the gap between current AI programs, which are
narrow in scope, and the types of AGI systems commonly seen in fiction” (Goertzel, 2014,
p.2). Unlike ANI, AGI systems will possess the ability to understand or learn any intellec-
tual task that we can. It will also have the ability to reason. Such systems would have a more
general-purpose intelligence and not be restricted to a specific task. AGI is considered the
next step in the evolution of AI and is often referred to as “strong AI” (Mills, 2018).
According to research and news articles, AI research labs and companies have not reached
this stage of development (Baum & Villasenor, 2023; Bove, 2023; Gillani, 2023), but AGI is
the goal of many researchers in the field of AI. An example of AGI could be a tool that is
able to initiate and perform a wide range of tasks such as a teaching software that can
assign classwork, devise schedules for students, correct classwork, coordinate with admin-
istrative staff, community partners, and other students, and adapt to new situations, all
without the need for human intervention or reprogramming.
Artificial Super Intelligence (ASI) is the final conceptualised type of AI evolution
(Yampolskiy, 2016), where systems would surpass human intelligence in every possible
aspect and possess the ability to outsmart us in all tasks, including higher-order thinking
skills such as creativity, general knowledge, and problem solving. Some believe that ASI
systems would have the ability to solve complex problems, make decisions, and even possess
consciousness and emotions (e.g., Barney, 2023; Marri, 2018). This type of sophistication
is sometimes referred to as “AI singularity” (Grossman, 2023). The term “singularity” in
this context is borrowed from the field of physics, where it is used to describe a point in
space-time where the rules, as we understand them, break down, such as at the centre of a
black hole. In the context of AI, the singularity would represent a point where machines
become so advanced that human minds can no longer predict or understand the former’s
behaviour or implications. This could include AI improving its design and creating succes-
sive generations of increasingly powerful machines. This type of AI is still purely specula-
tive, and the exact timeline for its development is unknown. However, the development of
ASI and its impact on society raises many ethical and moral questions. ASI is often por-
trayed as a potential threat to humanity, and there are debates about how to ensure its safe
development and deployment (Bostrom, 2014; Yudkowsky, 2006). When contemplating
ASI, it is often associated with fictional depictions such as Skynet in The Terminator, Vision
from the Marvel franchise, or the AI robot (Sonny) in I, Robot.
Currently, we are still in the stage of ANI, where our systems are designed to perform
specific tasks, whereas AGI and ASI remain as areas of continued research and specula-
tion. The potential for future advancements is immense, and the impact that AI will have
on society continues to be the subject of research and debate.

1.3 Big Data and the Internet of Things (IoT)

The growth of Big Data and the Internet of Things (IoT) has created a platform in which
AI can be applied in a wide range of industries and applications, from healthcare and
finance to retail and transportation. The progress of Big Data and IoT also greatly impacts
how we live, enabling greater convenience and efficiency, in turn leading to improved pro-
ductivity and enhanced quality of life.
4 Introduction to Artificial Intelligence in Higher Education
1.3.1 What Is Big Data?

You may have heard of the term “Big Data”. As indicated by its name, Big Data refers to a
vast amount of data (Favaretto et al., 2020; Kitchin & McArdle, 2016), but did you know
that data can come in various forms, such as text, numbers, images, audio recordings,
and videos?
Big Data refers to extremely large datasets that are too complex and voluminous for
traditional data processing systems to handle efficiently (McAfee & Brynjolfsson, 2012).
They can be structured, semi-structured, and unstructured data generated or retrieved
from various sources, including social media, e-commerce platforms, sensors, government
data, financial data, telecommunication and gaming industries, and system and software
logs. Structured data is information that is organised in a specific and predefined format,
such as tables, spreadsheets, and databases (Sarker, 2021). The data is already somewhat
categorised. Examples of structured data include customer records, financial transactions,
and inventory lists. This type of data is easily searchable, sortable, and analysed using con-
ventional data processing tools. On the other hand, unstructured data is information that
does not fit neatly into predefined categories or structures (Sarker, 2021). It often requires
specialised tools and techniques to extract meaningful insights. Examples of unstructured
data include social media posts, emails, and customer feedback.
Different types of data can enhance the comprehensiveness and precision of data anal-
ysis. The availability of large amounts of data has allowed AI algorithms to learn and make
predictions more accurately. With Big Data, AI models can be trained on larger datasets
and make predictions that are more informed and reliable. As a result, managing, process-
ing, and extracting insights from Big Data often requires specialised tools and techniques,
for example data warehousing, machine learning, and distributed computing. Yet, it is not
only the amount of data that is important – it is what organisations do with the data. Big
Data can be analysed for insights which lead to better decisions and strategic moves.
Table 1.1 shows a list of examples of various sources of Big Data (Agrawal et al., 2011;
Raghupathi & Raghupathi, 2014; Sagiroglu & Sinanc, 2013).

1.3.1.1 Characteristics of Big Data

In 2001, Gartner analyst Doug Laney first introduced “three Vs” as the main traits of Big
Data – Volume, Velocity, and Variety (Laney, 2001). These “three Vs” form a key framework
which influences how organisations and researchers approach and use Big Data. In addition
to Laney’s work, studies have suggested other significant characteristics, such as Veracity,
Variability, and Value (Jin et al., 2015; Kitchin & McArdle, 2016; Sagiroglu & Sinanc, 2013;
Sharma, 2022; Song & Zhu, 2015). These characteristics, summarised in Table 1.2, are often
used to support organisations in making decisions regarding Big Data, and to determine the
best ways to process, analyse, as well as extract insights from said data. Kitchin and McArdle
(2016) provided a clear overview of the ontological traits of Big Data.

1.3.1.2 Example of Big Data and Its Characteristics in Higher Education

Big Data in education – for example, educational data and analytics regarding student
learning – has all of the above characteristics. The volume can refer to collecting and ana-
lysing large amounts of data from various sources, including student transcripts, test scores,
 Introduction to Artificial Intelligence in Higher Education 5
Table 1.1 Examples of Various Sources of Big Data
Social Media
• Social network posts: Posts made on social media platforms such as Meta (Facebook), X
(previously Twitter), Instagram, and LinkedIn.
• Online videos: Videos uploaded to platforms like YouTube, Vimeo, and others.
• User-generated content: Reviews, ratings, and other types of content generated by users on
various websites and platforms.
• Social media advertising: Data collected by advertisers on users’ behaviours, preferences, and
responses to ads on social media platforms.
E-commerce
• E-commerce transactions: Online purchases made through websites such as Amazon, eBay,
and others.
• Banking transactions: Transactions made through online banking platforms, including bill
payments, money transfers, and other financial transactions.
• Stock market transactions: Transactions made through online stock trading platforms.
• Healthcare transactions: Electronic health record (EHR) transactions and other healthcare-
related transactions made through online platforms.
Sensors
• Environmental sensors: Tracked records of temperature, humidity, air pressure, and other
environmental parameters.
• GPS sensors: Tracked locations and movements of people, vehicles, and other objects.
• Medical sensors: Used in healthcare to monitor vital signs and track patients’ health status.
• Industrial sensors: Monitored performance of industrial machinery and tracked production
processes.
• Traffic sensors: Tracked flow of vehicles on roads and highways used to provide data for
traffic management and analysis.
• Retail sensors: Monitored customer behaviour in retail environments used to provide data for
marketing and sales analysis.
• Home automation sensors: Tracked usage records of appliances, lighting, heating and
cooling, and other systems in smart homes.
Government, Financial, Telecommunication and Gaming Industries
• Government: Data generated by government agencies, such as demographic data, crime
statistics, and tax records.
• Financial services: Data generated by financial institutions, including transaction records,
customer behaviour, and market data.
• Telecommunications services: Data generated by telecommunication companies, including call
records, network performance data, and customer usage data.
• Gaming companies: Data generated by online gaming platforms, including player behaviour,
in-game transactions, and performance metrics.
System and Software Logs
• Web server logs: Tracked records of all the requests made to a web server, including
information such as the type of request, the source IP address, and the time of the request.
• Application logs: Tracked records of events generated by software applications, such as error
messages, performance metrics, and user actions.
• Network logs: Tracked records of network activity, including information about network
traffic, security events, and system events.
• System logs: Tracked records of events generated by operating systems, such as boot events,
system crashes, and security events.
• Security logs: Tracked records of security-related events, such as attempted and successful
login attempts, changes to security settings, and security incidents.
6 Introduction to Artificial Intelligence in Higher Education
Table 1.2 Characteristics of Big Data
Volume refers to the large amounts of data generated and stored by organisations, individuals, and
systems. With the rise of digital technologies, Big Data often involves petabytes, exabytes, or
even zettabytes of data.
Velocity refers to the speed at which data is generated and continuously processed. Big Data is
generated and processed in real time, requiring us to respond quickly and use large amounts of
processing power to keep up with changing circumstances.
Variety refers to the diverse types and formats of data present within Big Data. These can range
from structured data, such as spreadsheets or databases, to unstructured data, such as texts
from social media commentaries.
Veracity, also known as data assurance, refers to the quality, accuracy, and integrity of the data
(Gillis, 2021; Raghupathi & Raghupathi, 2014). This is an important aspect as Big Data is only
useful if the data is reliable and trustworthy. Examples include the completeness of data or data
entry errors from typos and incorrect entries.
Variability refers to the meaning of data constantly changing (McNulty, 2017). For example, the
interpretation of certain words may depend on the specific context(s) in which they are used.
Value refers to what can be done with data, and the value(s) that Big Data can offer (Gillis, 2021).
How value is inferred from data is distinct from case to case.
Complexity refers to complications in a dataset, such as data with time-varying covariates from
scientific simulations (Hong et al., 2018). Data can be very complex in structure and may have
complicated relationships, making it difficult to identify patterns, correlations, and insights. An
example from the educational setting that demonstrates the complexity of Big Data is analysing
large amounts of data from various sources for student retention. There may be multiple
variables at play, such as individual student characteristics, academic support services, and
pedagogies of institutions, which should be considered within the data analysis. The complexity
of Big Data in this context requires advanced analytical skills, powerful data processing tools,
and a robust data management infrastructure.
Exhaustivity refers to the comprehensiveness of the data being analysed (Mayer-Schonberger &
Cukier, 2013). It signifies that the data should cover all aspects of the topic being analysed to
provide a complete picture. An example is student performance, where exhaustiveness is
achieved by including the entire population of students and not just capturing a sample, and
that data is gathered from all related aspects, for example test scores, tutorial and homework
completion rates, class participation levels, extra-curricular activities, and achievements, as well
as teacher evaluations.
Relationality refers to the relationship and connections between datasets and the ways in which
they can be linked, related, and integrated to provide meaningful insights and information
(Boyd & Crawford, 2012). An example from the financial industry is where large amounts of
data from various sources such as the stock market, information on customer behaviours, and
market trends are integrated for analysis to identify trends and correlations.
Extensionality refers to how data can be extended, expanded, or integrated with other data
sources and systems (Marz & Warren, 2012). This includes the ability to add new data sources,
and to combine or relate data from different sources to achieve a more comprehensive view of
the information collected. For example, health records could be analysed with data generated
from clinical research and healthcare devices, offering more comprehensive insights for the
identification of disease patterns, as well as advance patient care and treatment.

demographic information, and attendance and engagement data from online learning plat-
forms. The velocity of the data could refer to real-time updates of student progress and
achievements on a learning platform. The variety of data sources collected can be vast,
including grades, scores from assignments (such as reflective journals, video presentations,
and portfolios), and data logs of student activity (such as logs from communication chan-
nels and attendance records). Veracity of the data may refer to students giving incorrect
information with errors and typos, or missing data as students drop courses after a certain
 Introduction to Artificial Intelligence in Higher Education 7
period of time. To address these challenges, institutions may use complex algorithms and
models to extract insights from collected data, and make informed decisions for areas
including student retention, curriculum design, and resource allocation. Integrating and
analysing data from various sources and in different formats using innovative models may
contribute to the complexity of the Big Data. Variability is another a concern, such as stu-
dents having different learning paces and patterns. The exhaustiveness of the data is also a
challenge, as the data collected may not capture the full picture of student performance. As
an example, factors such as learning environment or socio-economic background which
may influence student learning could be overlooked. The relationality of the data, such as
linking students’ grades to the time of day and to the courses they have taken, can enable
the system to provide meaningful insights. Finally, the extensionality of Big Data may refer
to collection of data on students who fail, change programmes, undertake an overseas
exchange, or take a gap year, to gain a more comprehensive understanding of student
learning.

1.3.2 What is the Internet of Things (IoT)?

The Internet of Things (IoT) refers to the network of physical objects and things (e.g.,
smartphones, smart watches, home automation controls, fitness trackers) which can inter-
act and cooperate with each other through wired and wireless networks, as well as generate,
exchange, send, and receive data (Atzori et al., 2010; Li et al., 2015). IoT devices generate
vast amounts of data which, when combined with Big Data, can be used to train and
improve AI algorithms. IoT and AI are highly complementary technologies, with AI capa-
ble of analysing the volumes of data generated by IoT devices, including real-time data, in
order to provide valuable insights and facilitate prompt and automated decision-making.
IoT could also be a game changer for sustainability: according to an article from the World
Economic Forum (Arias et al., 2018), 84% of existing IoT deployments have the ability to
contribute to the achievement of the United Nations’ Sustainable Development Goals. IoT
provides a platform to connect everything and everyone in a smarter way, improving our
quality of life and benefitting our planet.
Those unfamiliar with IoT may hold a misconception that it only refers to devices con-
nected to the Internet; however, IoT encompasses much more than that. Devices with elec-
tronics parts, software, and sensors that allow them to connect and exchange data, such as
Bluetooth devices, are also part of IoT. The term “Internet of Things” was coined by Kevin
Ashton in 1999 during his employment at Procter & Gamble (Lueth, 2014; Ashton, 2009).
During his work to optimise supply chains, he sought to bring to his senior management’s
attention, a promising technology known as Radio Frequency Identification – RFID. Given
the widespread popularity of the internet at the time and its relevance to the concept,
Ashton labelled his presentation with the title, “Internet of Things”. However, it was only
since 2010 that IoT began gaining widespread attention (Lueth, 2014).

1.3.2.1 Characteristics of IoT

IoT has distinct characteristics. By understanding what they are, one can better grasp and
identify the underlying technology of IoT and its capabilities, determine the scope of IoT
applications, the types of devices and systems that IoT incorporates and can be integrated
with, its limitations, and the security and privacy implications. This information can be
8 Introduction to Artificial Intelligence in Higher Education
Table 1.3 Characteristics of IoT
Connectivity: IoT refers to the network between two or more objects or things which can
communicate and exchange data, such as the connection of a device to the internet.
Interoperability: IoT devices need to work together seamlessly regardless of their underlying
technology, platform, or brand names (Magan, 2023). Interoperability between devices is
essential to ensure that data can be collected, processed, and analysed effectively.
Scalability: IoT can be scaled, where systems and networks can be easily expanded or reduced in
size to adapt to changing needs, without affecting their performance (Magan, 2023). Scalability
is critical to ensure that IoT systems and networks remain functional and efficient as the number
of connected devices increases.
Intelligence: IoT devices, systems, and networks must have the “intelligence” to sense and analyse
data and interact with each other, using a combination of algorithms, software, and hardware.
Identity: Identity is a concept of IoT (Uikey, 2022) where, for example, the unique identity of an
IoT device enables the tracking and remote control of this specific device (e.g., monitoring
cameras installed in homes).
Autonomous: IoT systems have the ability to operate independently, without the need for human
intervention.
Security: IoT systems and devices require proper, built-in security measures given their
vulnerability to cyber threats. IoT devices often hold sensitive personal information; if the
device is lost or stolen, it may cause safety issues for the user. Firewalls, encryption, data
protection software, and access controls can be integrated into IoT devices to prevent data theft
and unauthorised access.
Real-time: IoT systems and devices enable real-time connection and data exchange.
Remote monitoring: IoT systems have monitoring capabilities. With remote monitoring, IoT
sensors can allow transmission of precise data for monitoring of device conditions and control
over device performance (Murphy, 2022).
Standardisation: IoT devices should follow standardised protocols, data formats, and rules to
ensure interoperability, consistency, and ethical use. Examples of protocols and standards
include the Constrained Application Protocol (CoAP), Message Queuing Telemetry Transport
(MQTT), and Bluetooth Low Energy (BLE) (Pratt, 2023).
Sustainability: IoT networks, systems, and devices consume energy and generate waste. Therefore,
it is important to consider their impact on the environment and to ensure that they are
sustainable.

used to make informed decisions about the design, deployment, and use of IoT systems
and devices. Furthermore, a profound understanding of IoT characteristics can facilitate
development of innovative IoT solutions and implications which can address and over-
come the challenges and limitations of existing systems and devices (Motta et al., 2019;
Rose et al., 2015; Atzori et al., 2010). IoT is characterised by the key features given in
Table 1.3.

1.3.2.2 Example of IoT and Its Characteristics in Higher Education

A smart classroom integrating different IoT systems and devices, such as sensors, cameras,
laptops, tablets, interactive whiteboards, learning management systems, student informa-
tion systems, printers, scanners, and wearable devices, is an example of how IoT can be
implemented in education.
Connectivity enables devices and systems to form a network for transmitting and receiv-
ing data as required. IoT in the classroom is scalable, capable of changing the scope of its
 Introduction to Artificial Intelligence in Higher Education 9
tasks and easily accommodating the number of devices connected. Each device in the net-
work has a unique identity, ensuring that students and teachers can securely identify and
access the devices and data they need. Interoperability enables seamless streamlining pro-
cesses such as in the submission of assignments and sharing of documents or resources
between students and teachers. Security is particularly important in a smart classroom as it
is essential to protect data, especially sensitive information such as grades and students’
personal details, from unauthorised access. IoT devices in a smart classroom must be
designed with robust security measures to ensure the protection of private and sensitive
information. The devices in the network are also autonomous, meaning they can make deci-
sions and take actions without human intervention. For example, if a room becomes too
warm, a smart sensor can automatically adjust the temperature. The system further has
intelligence capabilities such as machine learning algorithms, to process collected data and
make predictions to benefit the class. This enables teachers to track student progress and
identify areas where additional support is needed. The system follows industry standardisa-
tion to ensure that different devices from different vendors can work effectively and seam-
lessly together under standardised protocols. Additionally, the devices and systems used are
designed for sustainability, utilising energy-efficient and waste-reduction technologies to
reduce the carbon footprint, while maintaining cost-effectiveness with automatic lighting
and air conditioning systems. By incorporating these characteristics, this smart classroom
system example provides a solution for higher education institutions to improve student
experiences, streamline operations with technologies, and create a more efficient and sus-
tainable learning environment.

1.4 Generative Artificial Intelligence (GenAI)

Generative AI (GenAI) is a subset of artificial intelligence that involves creating new data
or content (Dilmegani, 2023; McKinsey Consultant, 2023; Altexsoft, 2022). The term
“generative” refers to the capability of AI to produce novel outputs rather than simply
reproducing, categorising, processing, and analysing inputs. It can generate anything from
text, images, audio, and videos, to natural-language text and even computer coding.
The concept of GenAI is rooted in the field of computer science and artificial intelli-
gence. One of the earliest references to such a technology can be found in the work of
computer scientist John Hopfield in the 1980s, who proposed the idea of a “Hopfield net-
work”, a type of computer neural network that could be trained to generate new patterns
or sequences (Kampakis, 2022; McKinsey Global Institute, 2018). The fields of machine
learning and artificial neural networks have evolved over time as researchers and practi-
tioners worked to develop more sophisticated generative architecture and deep learning
algorithms, including generative adversarial networks (GANs), variational autoencoders
(VAEs), and transformer-based models (e.g. GPT-3 Generative Pretrained Transformer 3).
There are two important models in GenAI: Discriminative and Generative Modelling
(Altexsoft, 2022).
Discriminative modelling is a type of machine learning approach that aims to classify or
categorise existing data points into specific classes or categories. While this model is able to
distinguish the boundaries between classes, it is trained only to tell the differences and does
not understand how data is generated. Machine learning tasks using this approach are
classified as supervised machine learning tasks. The goal of this approach is to discriminate
10 Introduction to Artificial Intelligence in Higher Education
or make a distinction between the different classes, and then make predictions based on the
data it has been trained on.
For example, in education, let’s say we want to develop a model that can predict whether
a student will pass or fail a particular exam based on their past performance, test scores,
and attendance records. The discriminative model would look at the data provided, identify
the differences between students who passed the exam and those who failed, then make
predictions based on those differences. The model is simply trying to distinguish between
students who passed and those who failed – it does not consider why students achieve cer-
tain outcomes, and only aims to make accurate predictions.
Generative modelling is a type of machine learning approach that focuses on creating
new data that resembles the data it was trained on, based on the patterns and relationships
learned from existing data. Machine learning tasks using this approach are classified as
unsupervised and semi-supervised machine learning tasks. The goal of the generative model-
ling approach is to create new examples that are similar to those it has seen before, which
are representative of the underlying patterns and relationships in the data.
An example of generative modelling in education could be a system that generates new
questions for a quiz or test. The model would learn the patterns and distribution of past
questions from a training set, and subsequently use that information to generate new ques-
tions that are similar in style and content. This could be useful for creating new practice
problems for a class, adaptive tests for each student based on their learning progress, or
creating unique tests for individual students to prevent cheating.

1.4.1 Applications of GenAI

Today, GenAI is already widely applied across various industries, including image and
audio synthesis, and text, code and video generation. For example, GenAI can be used to
generate realistic images of objects or scenes which do not actually exist, or to create syn-
thetic speech that sounds like a real person. GenAI can also be used to generate new designs
or prototypes in fields such as engineering and architecture, or to generate new music com-
positions or pieces of art (Kumar, 2023; Marr, 2023). Figure 1.2 shows an example of an
image synthesis and Table 1.4 shows an example of a code generation. Table 1.5 provides
some examples of the types of applications and software of GenAI.
Most applications use a combination of image, text, code, video, and audio generation
alongside machine learning algorithms to produce solutions to a problem. One example of
how all of these could be used in a comprehensive educational tool could be through an
intelligent tutoring system. This system can use GenAI algorithms to customise learning
materials for individual students based on their learning needs and progress, produce per-
sonalised video lessons by narrating snippets of educational text or code, combine it with
visual aids such as images or animations, in order to support students in understanding
complex concepts. The system can also create assessment and grading descriptors to assess
students’ understanding of materials, generating questions and quizzes to reinforce learn-
ing. Moreover, audio feedback can also be generated for students as based on their perfor-
mance, providing guidance and encouragement in a natural and engaging manner. This
example demonstrates the transformative capabilities of GenAI in education to create an
immersive, personalised, and effective learning experience. By harnessing the power of AI,
teachers can free up time and resources, allowing them to focus on other priorities like
classroom instruction and providing support to students.
 Introduction to Artificial Intelligence in Higher Education 11
User You are VanGoghBot. You will generate painngs in the style of Van Gogh.
Of course! I'm ready to create painngs in the style of Van Gogh. Please
ChatGPT
describe the scenes or subjects you'd like to see, and I'll generate painngs
DALL.E3
based on your descripons.
User Please draw me an electrical car with a robot

ChatGPT
DALL. E3

Figure 1.2 ChatGPT Dall.E3 Generated Images.

1.5 ChatGPT – A Game Changer in Education?

One particular AI tool that has gained massive popularity, including in education, is
ChatGPT (Chat Generative Pre-trained Transformer). ChatGPT is a pretrained text-­
generation language model developed by OpenAI that uses a transformer-based language
architecture (Srivastava, 2022; Joshi, 2021; Kulshrestha, 2020), a type of deep neural net-
work well suited for natural language processing tasks. It has the ability to understand and
generate human-like text based on the input it receives. In addition, it has been trained on
a massive amount of data which has allowed it to learn the patterns, styles, and complexi-
ties of human language, making it a remarkable tool for communication. With capabilities
in language generation, translation, summarisation, and questioning and answering,
ChatGPT has caused a huge ripple in the technological industry (Zaveri, 2023).
The rise of ChatGPT and its popularity are connected to a number of factors: ChatGPT
is redefining the boundaries of machine learning (Rosenbaum, 2023); it is efficient in gen-
erating results with endless possibilities (Pace, 2023; Petrova, 2023); and it is knowledgeable
with training from huge volumes of information (Shankland, 2023). AI has been around
for more than a decade, but recent advancements in the field have made it more accessible
12 Introduction to Artificial Intelligence in Higher Education
Table 1.4 ChatGPT Generated Codes
User Can you create a plugin for a simulated DoctorBot?
ChatGPT Certainly! Creating a plugin for a simulated DoctorBot involves multiple steps,
including defining the purpose, designing the user interface, and implementing the
logic. Below is a basic outline for creating a DoctorBot plugin:
1. Define the Purpose:
Symptom checker/General medical information/Medication reminders etc.
2. Design the User Interface:
Text-based chat interface/Voice recognition/Multi-choice questions etc.
3. Implement the Logic:
For simplicity, let’s create a basic symptom checker using Python:
class DoctorBot:
def __init__(self):
self.symptoms = {
“headache”: “You might have a migraine or tension headache. If it
persists, please see a doctor.”,
“fever”: “You might have an infection. Please monitor your temperature
and see a doctor if it remains high.”,
“cough”: “It could be a cold or flu. If it persists or is severe, please see a
doctor.”
}
def check_symptom(self, symptom):
return self.symptoms.get(symptom, “I’m not sure about that symptom.
Please consult a doctor.”)
def main():
bot = DoctorBot()
while True:
symptom = input(“Please describe your symptom: “).lower()
if symptom == “exit”:
break
print(bot.check_symptom(symptom))
if __name__ == “__main__”:
main()

and user-friendly. With ChatGPT, OpenAI’s decision to allow public access to the software,
its user-friendly interface, and its ability to understand and respond to natural language in
a human-like manner, sets it apart from previous AI models. Additionally, the COVID-19
pandemic has accelerated the shift towards online and virtual interactions. The public has
grown in their awareness and acceptance of online functionality given their increased expe-
riences with remote working and learning, as well as through the increased popularity of
online financing and shopping. As a result, there is a greater demand for innovative solu-
tions like ChatGPT to cater to different needs of the public (Zaveri, 2023).
Within the field of education, the pandemic has become a stimulus for individuals to
recognise the benefits and potentials of full-time virtual learning. Before the pandemic,
traditional, in-person teaching and learning was regarded as the sole, favoured, or most
effective approach to education for many individuals. The sudden outbreak and spread of
COVID-19, resulting in widespread lockdowns and social distancing measures, forced uni-
versities, colleges, and schools to close their physical campuses and transition to virtual
teaching and learning. With the sudden shift online, many have had to adapt and embrace
online education, leading to gradual recognition of the possibilities and capabilities such a
 Introduction to Artificial Intelligence in Higher Education 13
Table 1.5 Examples of Types of Generative AI Applications and Software
Applications Industries/Purposes Popular Software Apps

Image synthesis: Generative AI Art and Design, Stable Diffusion, Dall-E,

algorithms can be used to create Healthcare, Film, Runway ML, Midjourney,
unique and creative works of art, such Advertising DeepAI, Fotor, Reface,
as digital paintings, sculptures, and FaceApp, SpeakPic,
graphic designs. It can also be used in Syntegra, AlphaFold2
medical imaging analysis to detect
early signs of diseases and medical
anomalies. It is important to mention
Deepfakes, a type of AI technology
that allows the generation of synthetic
images and videos of people. Image
recognition and processing in AI
often use a Convolutional Neural
Network (CNN).
Video generation: Generative AI can be Film, Advertising, Pictory, Synthesia, VEED
used to create videos that look like Customer Service for
they were shot in real life, or create an various industries
animated short film.
Music and audio creation: Generative AI Music Production, Amper Music, Jukebox,
can be used to create new music Healthcare, Customer MuseTree, MuseNet,
compositions, remix existing songs, Service for various Musico, Vall-E, WaveNet,
and generate custom sound effects. industries Flowtron
Text-to-speech, speech-to-speech: Customer service for Siri, Alexa, Google
Generative AI can be used to various industries, Assistant
convert text into speech or facilitate Electronics, Travel,
real-time speech translation. For Leisure
example, it can enable virtual
assistants to “speak” in applications
or devices.
Text generation: Generative AI can be Marketing, Advertising, ChatGPT, BARD, Bing,
used to create copywriting, Journalism, Script and Quillbot, Jasper, Google
advertisements, and content for Story Writing, Translate, Photomath
marketing campaigns. It can be used Healthcare, Customer
to write stories, articles, and other Service for various
forms of text, including conversations industries
with virtual assistants and help
chatbots. It also has the power to
translate text from one language to
another, breaking down language
barriers and enabling more efficient
communication.
Code generation: Generative AI can All industries that require Google’s AutoML, OpenAI
develop, write, and debug computer programming Codex, GitHub Co-pilot,
programming codes. Polycoder, Amazon
CodeWhisperer

mode holds for delivering high-quality education. This has also opened up new opportuni-
ties for utilising technology in classrooms. For example, ChatGPT is expected to play a
significant role in the future of teaching and learning, driving education to develop and
undertake better practices by replacing those which could be automated (Jacobsen, 2023).
In the post-COVID era, we are beginning to see many educational institutions opening up
14 Introduction to Artificial Intelligence in Higher Education
to the possibility of allowing students to continue their education remotely, which increases
accessibility and flexibility, as well as creates new opportunities for learning. Even renowned
universities such as University College London (UCL) or Kings College London are pro-
viding numerous pathways to go remote by offering online Masters programmes (UCL,
n.d.) and blended options for International Foundation programmes (King’s College
London, n.d.). Online courses allow students to attend lectures and complete assignments
from the comfort of their own homes, which can save time and money on commuting.
Moreover, virtual learning environments can be designed to promote inclusivity, better
support students with special needs, and ensure that those in remote areas can still access
teaching and learning. The growing demand for online learning has also led to the increased
popularity and adoption of AI-powered virtual assistants like ChatGPT, which can further
enhance the accessibility, efficiency, and effectiveness of education. For instance, ChatGPT
can serve as a 24/7 virtual assistant that offers real-time responses to students’ queries,
provide personalised support and guidance, and provide assistance in maintaining learning
progress.
All in all, the COVID-19 pandemic has accelerated the shift towards online and virtual
learning. As a result, behaviours and attitudes of the public have changed, and educational
audiences are now much more receptive to the use of AI tools. We should also note that the
recent surge of discussion regarding AI’s impact on education is not necessarily late; despite
the 50-year presence of AI, its impact on education has only gained considerable promi-
nence with the recent advancement of ChatGPT. While tools like Google Translate may
have made notable impacts on certain educational domains, its impact pales in comparison
to the potentials of ChatGPT. With its capabilities of providing immediate and tailored
feedback, facilitating natural language interactions, and serving as both a teaching and
research assistant, ChatGPT and other GenAI tools have established themselves as a trans-
formative force for higher education. Such technology has the potential to fundamentally
change the way students learn while also supporting educators in overcoming challenges
and obstacles in teaching and learning.

1.6 Generative AI Concerns and Challenges

Despite its potential benefits, the utilisation of AI raises important ethical and social
issues. These include concerns about job displacement, privacy, and accountability, as
well as the potential for AI to be used for malicious purposes. One of the key drivers
behind the development of generative AI is the desire to elevate the creative process
(Eapen et al., 2023). By implementing algorithms and machine learning models, genera-
tive AI possesses the capacity to generate new and unique content that cannot be achieved
through traditional, manual methods. However, the benefit of producing novel outputs
also introduces potential drawbacks, including the risk of generating biased or unfair
outputs, or content that is inappropriate or harmful. Several aspects are to be addressed:
the need to strike a balance between human and machine values and control; the need to
ensure that generated content is of high quality and relevancy; and the need to navigate
the ethical considerations arising from the complex nature of AI. It is important that we
carefully consider these issues as we continue to develop and deploy AI in across indus-
tries. In Chapter 3, we will explore and further discuss opportunities and concerns for
GenAI in higher education.
 Introduction to Artificial Intelligence in Higher Education 15

GPT
Input •Encoding Output
(Prompt) •Aenon: word weights (Response)
•Decoding: Predicon Sequence

Figure 1.3 How Does ChatGPT Work?

1.7 How Does ChatGPT Work?

For many readers, they are interested in understanding the big picture how ChatGPT
works, thus, we have put together a high-level overview of how it works below, please also
see Figure 1.3. If you wish to go deeper into the technicalities, please read Chapter 6.
In the simplest way of explaining, ChatGPT is a type of software trained on huge
amounts of text. During its training phase, ChatGPT is exposed to vast amounts of text
from variety of sources such as websites, books, forums. It learns by predicting the next
word in sentences. The training helps ChatGPT to recognise patterns and structures in
language. It uses a structure called the Transformer to weigh which parts of your question
are most important and how your prompt relates to all the data it has trained on, and thus,
GPT stands for Generative Pre-trained Transformer.

1.7.1 So What Happens When You Type in a Question (Also Known as Prompt)?

Table 1.6 shows the steps of how ChatGPT works.

Input: You type in a sequence of words (e.g., “What is the capital of ”) also known as
a prompt.

Table 1.6 Steps Showing How ChatGPT Works

What happens when you type in a question (prompt) in ChatGPT

[User Input/Prompt in → “What is the capital of ?”

ChatGPT]
[Encoding] → Vector for ‘capital’ for example might relate to a every token
form its training that is a capital e.g. ‘Dublin’, ‘London’,
‘Paris’, etc. The vector for ‘Paris’ which already exists from
training would relate to ‘France’
→ Vector for the other words are created in a similar way
[Attention Mechanism: [Attention Mechanism: Word Weights]
Word Weights] → “What” (medium weight)
→ “is” (low weight)
→ “the” (medium weight)
→ “capital” (high weight)
→ “of ” (medium weight)
[Decoding] → ChatGPT grasps that the user is asking about a capital city
[Prediction Sequence] → “The capital of ”
→ “The capital of France”
→ “The capital of France is”
→ “The capital of France is Paris.”
[ChatGPT Output/Response] → “The capital of France is Paris.”
16 Introduction to Artificial Intelligence in Higher Education
Encoding: The model translates each word into a numerical representation called a vector.
This representation captures the word’s meaning and its relation to other words which it
has been trained on.
Attention: Using a method called “self-attention”, the ChatGPT model weighs the
importance of each word in your question. For instance, in our example, “capital”
might be given more weight than “is”. The goal is to grasp the essence of your ques-
tion/prompt. Knowing the topic helps the model understand related words, like “of ”
and “the”.
Decoding & Prediction: Once the model understands the context of your question from
the word weightings (attention) and relationship to other words it has been trained on
(embedding), it starts predicting the best sequence of words for a response. This is
based on the input – the encoding information and the attention weights. GPT pro-
duces a series of words as an answer. Each word has a specific relevance to the topic and
the other words. The model predicts the next word. Then it continues, using both the
original input and the words it has already predicted to guess subsequent words. This
process repeats until a full answer is formed or until the model decides the answer is
complete.
Output: The model provides the generated sequence of words as a response (e.g., “The
capital of France is Paris.”).

It is important to note that while this explanation provides an overview, the actual work-
ings of the model involve a lot of intricate math and processes happening simultaneously.
The model has been trained on billions of such sequences, refining its ability to make accu-
rate and coherent predictions over time.

1.8 How does GenAI Generate Images from Text?

For many readers, the concept of machines creating images based on textual descriptions
might sound like science fiction. Yet, in GenAI, this has become a reality. To shed some
light on this capability, we have crafted a high-level overview of how GenAI works in the
context of generating images from text. Please refer to Figure 1.5 for a visual representa-
tion. Those desiring a deeper dive can explore Chapter 6.
Computers work with numbers. Whether it is text, image, or sound, a computer inter-
prets it as numbers. And every image is made of pixels. Each pixel has a colour, represented
by three numbers: red, green, and blue (RGB). An image, then, is essentially a matrix of
number trios. As you can see in Figure 1.4, digital images are represented by a size for
example 1920 pixels × 1080 pixels. Each pixel represents a location on the image like x, y
coordinates and then at each pixel there is a colour represented by three numbers for, one
for the amount of red, another for the amount of green, and another for the amount of
blue (RGB). This results in a matrix (array) of vectors (list of numbers) like [13,8,255,0,0]
which can represent the whole image.
Much like how ChatGPT encounters enormous amounts of textual data, GenAI Image
generation software is trained on vast visual and textual datasets. Think of it as an artist
with an encyclopedic knowledge of styles, subjects, and scenes. This can range from sim-
plistic drawings to intricate photographs.
Introduction to Artificial Intelligence in Higher Education 17
Figure 1.4 An Image Showing RGB Pixel.
18 Introduction to Artificial Intelligence in Higher Education

Text Encoder Image Generator

Input • Change to number • Use diffusion to Output
(Prompt) • Understand the create image (Response)
prompt

Figure 1.5 The Text to Image Process.

Through this training process, GenAI learns the complexity between words and visuals.
It begins to understand that the word “sunset” might evoke hues of orange and pink, or
that a “snowy mountain” corresponds to white peaks against a clear sky.
One approach to GenAI’s capability is the Generative Adversarial Network (GAN).
This involves a duo – a generator (the artist) and a discriminator (the critic). The artist
generates, and the critic evaluates. Over time, the generator gets better at producing images
that can pass the discriminator’s critique. Figure 1.5 shows how text is prompted and gen-
erated into image.

1.8.1 So What Happens When You Type in a Question (Also Known as Prompt) to
Generate an Image?

Input: A user provides a textual prompt (e.g., “A serene lake surrounded by autumn
trees”).
Textual Encoding: This process is similar to understanding text generation. To produce
an image from text, the model first encodes the text input into a meaningful numeri-
cal representation. This is usually done using a separate model, such as a trained
transformer.
Image Generator: Once the text is encoded into a numerical format, it is used as an input
or condition for the generator. This process uses diffusion to create the image.
Diffusion is a technique to make images fuzzy. This fuzzy appearance is similar to
static on old televisions, often referred to as “noise”. Noise is merely random colours/
pixels. You can make an image noisy by adding random numbers to its pixels.
Conversely, to clarify a noisy image, you adjust the random numbers so they produce
a coherent image. In this step, the generator then produces an image that corresponds
to the textual description as shown in Figure 1.6.
Refinement: The discriminator evaluates the generated image, ensuring its alignment
with the initial description. This feedback loop, over multiple iterations, refines the
visual output.
Output: The result is a visual representation (e.g., an image of a calm lake with trees
draped in autumnal colours) presented to the user as shown in Figure 1.7.

Figure 1.6 Showing the Diffusion Process.

 Introduction to Artificial Intelligence in Higher Education 19

The Image Response Generated From the Prompt “A Serene Lake Surrounded by
Figure 1.7 
Autumn Trees”.

1.9 Conclusions
The future of education is not just digital; it is smart, responsive, and tailored. Imagine
classrooms that function like a high-tech ecosystem, adjusting to the needs of every student
in real time. Imagine AI tools like ChatGPT, always ready to help, answering questions day
or night. COVID-19 pushed us online, and now, tech like IoT and GenAI is taking us to the
next level. The classroom of the future? It is innovative, flexible, and built on groundbreak-
ing technology. As we ride this wave, one thing is clear: learning is about to get a whole lot
more exciting.

Questions to Ponder

• What will happen if higher education institutions ignore the existence of GenAI?
• What will happen if higher education institutions embrace GenAI in teaching and
learning?
• What will happen if higher education institutions embrace GenAI in teaching and
learning, but do not put in place any relevant teaching and learning policies to
govern its implementation?
• Is ChatGPT really only an ANI? It has the capacity to generate new data, and
sometimes it seems to understand what we are asking beyond the prompt we enter.
• If ChatGPT is undertaking unsupervised or semi-supervised machine learning
tasks, how can we govern them to ensure certain biased, unethical tasks will not be
allowed; wouldn’t ChatGPT learn from its own generated content, and continue to
come up with and use its own new ideas and concepts – just like humans?
20 Introduction to Artificial Intelligence in Higher Education

Personal Reflection

If COVID-19 never happened, I wonder how we would perceive online learning and tools like
GenAI today? For me, I recalled implementing the e-learning policy at my university thirteen
years ago, and trying to get people on board with e-learning was a real struggle. Without
the pandemic forcing us to go online for work, school, and even socialising, would we still be
dragging our feet? Maybe the convenience of GenAI tools would have won us over eventually,
but certainly not as quickly. In a way, the pandemic forced us to dive into the deep end with
new tech. Without this happening, perhaps we would still be dipping our toes, hesitant to
fully plunge into digital learning. It is interesting to think about how necessity accelerates ac-
ceptance. What do you think? Would we have welcomed these smart tools without the push
from a global crisis? It is a curious path to wander down, imagining the “what ifs”.

References
Agrawal, D., Bernstein, P., Bertino, E., Davidson, S., Dayal, U., Franklin, M., Gehrke, J., Haas, L.,
Halevy, A., Han, J., Jagadish, H. V., Labridinis, A., Madden, S., Papakonstantinou, Y., Patel, J.,
Ramakrishnan, R., Ross, K., Shahabi, C., Suciu, D., … & Widom, J. (2011). Challenges and oppor-
tunities with Big Data 2011-1. [White paper]. Purdue e-Pubs. Retrieved October 17, 2023, from
https://docs.lib.purdue.edu/cctech/1/
Altexsoft. (2022, October 13). Generative AI models explained. Retrieved October 18, 2023, from
https://www.altexsoft.com/blog/generative-ai/
Arias, R, Lueth, K. L. & Rastogi, A. (2018, January 21). The effect of the Internet of Things on sus-
tainability. World Economic Forum. Retrieved 18 October, 2023, from https://www.weforum.org/
agenda/2018/01/effect-technology-sustainability-sdgs-internet-things-iot/
Ashton, K. (2009). That ‘Internet of Things’ thing. RFID Journal, 22(7), 97–114.
Atzori, L., Iera, A., & Morabito, G. (2010). The Internet of Things: A survey. Computer Networks,
54(15), 2787–2805. https://doi.org/10.1016/j.comnet.2010.05.010
Barney, N. (2023, March). Artificial Super Intelligence (ASI). TechTarget. Retrieved October 17, 2023,
from https://www.techtarget.com/searchenterpriseai/definition/artificial-superintelligence-ASI
Baum, J., & Villasenor, J. (2023, July 18). How close are we to AI that surpasses human intelligence?
Brookings. Retrieved October 17, 2023, from https://www.brookings.edu/articles/how-close-are-
we-to-ai-that-surpasses-human-intelligence/
Bostrom, N. (2014). Superintelligence: Paths, dangers, strategies. Oxford University Press.
Bove, T. (2023, May 4). CEO of Google’s DeepMind says we could be ‘just a few years’ from A.I. that
has human-level intelligence. Fortune. Retrieved October 17, 2023, from https://fortune.
com/2023/05/03/google-deepmind-ceo-agi-artificial-intelligence/
Boyd, D., & Crawford, K. (2012). Critical questions for big data: Provocations for a cultural, techno-
logical, and scholarly phenomenon. Information, Communication & Society, 15(5), 662–679.
https://doi.org/10.1080/1369118X.2012.678878
Cordeschi, R. (2007). AI turns fifty: Revising its origins. Applied Artificial Intelligence, 21(4–5), 259–
279. https://doi.org/10.1080/08839510701252304
Dilmegani, C. (2023, September 29). Top 100+ generative AI applications/use cases in 2023.
AIMultiple. Retrieved October 18, 2023, from https://research.aimultiple.com/generative-ai-
applications/
Eapen, T. T., Finkenstadt, D. J., Folk, J., & Venkataswamy, L. (2023). How generative AI can augment
human creativity. Harvard Business Review. Retrieved October 19, 2023, from https://hbr.
org/2023/07/how-generative-ai-can-augment-human-creativity
 Introduction to Artificial Intelligence in Higher Education 21
Favaretto, M., De Clercq, E., Schneble, C. O., Elger, B. S. (2020). What is your definition of Big Data?
Researchers’ understanding of the phenomenon of the decade. PLoS ONE, 15(2), e0228987.
https://doi.org/10.1371/journal.pone.0228987
Gillani, N. (2023, April 19). Will AI ever reach human-level intelligence? We asked five experts. The
Conversation. Retrieved October 19, 2023, from https://theconversation.com/will-ai-ever-reach-
human-level-intelligence-we-asked-five-experts-202515
Gillis, A. S. (2021, March). 5 V’s of Big Data. TechTarget. Retrieved October 18, 2023, from https://
www.techtarget.com/searchdatamanagement/definition/5-Vs-of-big-data
Goertzel, B. (2014). Artificial general intelligence: Concept, state of the art, and future prospects.
Journal of Artificial General Intelligence, 5(1), 1–46. https://doi.org/10.2478/jagi-2014-0001
Grossman, G. (2023, February 11). Generative AI may only be a foreshock to AI singularity.
VentureBeat. Retrieved October 17, 2023, from https://venturebeat.com/ai/generative-ai-may-only-
be-a-foreshock-to-ai-singularity/
Hong, Y., Zhang, M., & Meeker, W. Q. (2018). Big data and reliability applications: The complexity
dimension. Journal of Quality Technology, 50(2), 135–149. https://doi.org/10.1080/00224065.2018.
1438007
Jacobsen, P. (2023, February 8). Why ChatGPT will change Higher Ed for the better. FEE Stories.
Retrieved October 19, 2023, from https://fee.org/articles/chatgpt-will-change-higher-ed-for-the-better/
Jin, X., Wah, B. W., Cheng, X., & Wang, Y. (2015). Significance and challenges of big data research.
Big Data Research, 2(2), 59–64. https://doi.org/10.1016/j.bdr.2015.01.006
Joshi, P. (2021, July 23). Guide to the Latest State-of-the-Art Models. Analytics Vidhya. Retrieved
October 19, 2023, from https://www.analyticsvidhya.com/blog/2019/06/understanding-
transformers-nlp-state-of-the-art-models/
Kampakis, S. (2022, November 19). Tom Burns on Hopfield networks and the nature of general AI
[Interview]. Hackernoon. Retrieved October 18, 2023, from https://hackernoon.com/tom-burns-
on-hopfield-networks-and-the-nature-of-general-ai
King’s College London. (n.d.). King’s international foundation. Retrieved October 19, 2023, from
https://www.kcl.ac.uk/international-foundation
Kitchin, R., & McArdle, G. (2016). What makes Big Data, Big Data? Exploring the ontological char-
acteristics of 26 datasets. Big Data & Society, 3(1). https://doi.org/10.1177/2053951716631130
Kulshrestha, R. (2020, June 29). Transformers. Medium. Retrieved October 19, 2023, from https://
towardsdatascience.com/transformers-89034557de14
Kumar, S. M. (2023, June 23). How generative AI can transform architecture, engineering, construction.
The Economic Times. Retrieved October 19, 2023, from https://economictimes.indiatimes.com/
news/how-to/how-generative-ai-can-transform-architecture-engineering-construction/articles
how/101225327.cms?from=mdr
Laney, D. (2001). 3D data management: Controlling data volume, velocity and variety. META Group.
LeCun, Y, Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521, 436–444. https://doi.
org/10.1038/nature14539
Li, S., Xu, L. D. & Zhao, S. (2015). The Internet of Things: A survey. Information Systems Frontiers,
17, 243–259. https://doi.org/10.1007/s10796-014-9492-7
Lueth, K. L. (2014, December 19). Why the Internet of Things is called Internet of Things: Definition,
history, disambiguation. IoT Analytics Market Insights. Retrieved 18 October, 2023, from https://
iot-analytics.com/internet-of-things-definition/
Magan, S. (2023, October 18). Characteristics of IoT: Definition, uses & key features explained.
Unstop. Retrieved October 18, 2023, from https://unstop.com/blog/characteristics-of-iot
Marr, B. (2023, October 5). Generative AI is revolutionizing music: The vision for democratizing crea-
tion. Forbes. Retrieved October 19, 2023, from https://www.forbes.com/sites/bernardmarr/
2023/10/05/generative-ai-is-revolutionizing-music-loudlys-vision-for-democratizing-creation/?sh=
3c00b9b775bd
Marri, S. (2018, July 9). Can super intelligence and emotional intelligence co-exist? Forbes India.
Retrieved October 17, 2023, from https://www.forbesindia.com/blog/technology/can-super-
intelligence-and-emotional-intelligence-co-exist/
22 Introduction to Artificial Intelligence in Higher Education
Marz, N., & Warren, J. (2012). Big Data: Principles and best practices of scalable realtime data ­systems
(MEAP ed.). Manning.
Mayer-Schonberger, V., & Cukier, K. (2013). Big Data: A revolution that will change how we live, work
and think. John Murray.
McAfee, A., & Brynjolfsson, E. (2012). Big data: The management revolution. Harvard Business Review.
McKinsey Consultant. (2023, January 19). What is generative AI? Retrieved October 18, 2023, from
https://www.mckinsey.com/featured-insights/mckinsey-explainers/what-is-generative-ai
McKinsey Global Institute. (2018, May 8). Deep learning’s origins and pioneers. McKinsey &
Company. Retrieved October 18, 2023, from https://www.mckinsey.com/featured-insights/
artificial-intelligence/deep-learnings-origins-and-pioneers
McNulty, E, (2017, May 8). Understanding Big Data: The seven V’s. Dataconomy. Retrieved October
18, 2023, from http://dataconomy.com/seven-vs-big-data/
Mills, T. (2018, September 17). AI vs AGI: What’s the difference? Forbes. Retrieved October 17, 2023,
from https://www.forbes.com/sites/forbestechcouncil/2018/09/17/ai-vs-agi-whats-the-difference/?
sh=cc4150838ee1
Motta, R. C., Silva, V., & Travassos, G. H. (2019). Towards a more in-depth understanding of the IoT
Paradigm and its challenges. Journal of Software Engineering Research and Development, 7. https://
doi.org/10.5753/jserd.2019.14
Murphy, M. (2022, May 7). IoT remote monitoring: How it works, applications, benefits. Hologram.
Retrieved October 18, 2023, from https://www.hologram.io/blog/iot-remote-monitoring/
Pace, M. (2023, January 17). ChatGPT: Why the trending AI tool has captured attention and contro-
versy. Spinnaker. Retrieved October 19, 2023, from https://unfspinnaker.com/98982/features/
chatgpt-why-the-trending-ai-tool-has-captured-attention-and-controversy/
Petrova, M. (2023, February 2). Why ChatGPT is a game changer for AI [Video]. CNBC. https://www.
cnbc.com/video/2023/02/02/why-chatgpt-is-a-game-changer-for-ai.html
Pratt, M. K. (2023, July 12). Top 12 most commonly used IoT protocols and standards. TechTarget.
Retrieved October 18, 2023, from https://www.techtarget.com/iotagenda/tip/Top-12-most-
commonly-used-IoT-protocols-and-standards
Raghupathi, W., & Raghupathi, V. (2014). Big data analytics in healthcare: Promise and potential.
Health Information Science and Systems, 2. https://doi.org/10.1186/2047-2501-2-3
Rose, K., Eldridge, S., & Chapin, L. (2015). The internet of things: An overview. The Internet
Society (ISOC).
Rosenbaum, E. (2023, February 11). The ChatGPT AI hype cycle is peaking, but even tech skeptics
don’t expect a bust. CNBC. Retrieved from https://www.cnbc.com/2023/02/11/chatgpt-ai-hype-
cycle-is-peaking-but-even-tech-skeptics-doubt-a-bust.html
Sagiroglu, S., & Sinanc, S. (2013, May). Big data: A review. 2013 International Conference on
Collaboration Technologies and Systems (CTS), San Diego, USA (pp. 42–47). https://doi.
org/10.1109/CTS.2013.6567202
Sarker, I. H. (2021). Machine learning: Algorithms, real-world applications and research directions.
SN Computer Science, 2, 160. https://doi.org/10.1007/s42979-021-00592-x
Shankland, S. (2023, February 19). Why we’re obsessed with the mind-blowing ChatGPT AI chatbot.
CNET. Retrieved October 19, 2023, from https://www.cnet.com/tech/computing/why-were-all-
obsessed-with-the-mind-blowing-chatgpt-ai-chatbot/
Shao, Z., Zhao, R., Yuan, S., Ding, M., & Wang, Y. (2022). Tracing the evolution of AI in the past
decade and forecasting the emerging trends. Expert Systems with Applications, 209, 118211. https://
doi.org/10.1016/j.eswa.2022.118221
Sharma, R. (2022, September 1). Characteristics of Big Data: Types and 5V’s. UpGrad. Retrieved
October 17, 2023, from https://www.upgrad.com/blog/characteristics-of-big-data/
Song, I.-Y., & Zhu, Y. (2015). Big data and data science: What should we teach? Expert Systems,
33(4), 364–373. https://doi.org/10.1111/exsy.12130
Srivastava, A. (2022, August 8). What are transformers in NLP and its advantages. Knoldus. Retrieved
October 19, 2023, from https://blog.knoldus.com/what-are-transformers-in-nlp-and-its-advantages/
 Introduction to Artificial Intelligence in Higher Education 23
Thierer, A. D., O’Sullivan, A. O., & Russell, R. (2017). Artificial intelligence and public policy.
Mercatus Research, Mercatus Center at George Mason University, Arlington https://doi.
org/10.13140/RG.2.2.14942.33604
UCL (n.d.). Flexible learning for professionals – Online MSc programmes at UCL. UCL Faculty of
Medical Sciences. Retrieved October 19, 2023, from https://www.ucl.ac.uk/medical-sciences/study/
flexible-learning-professionals-online-msc-programmes-ucl
Uikey, P. (2022, November 26). Characteristics of IoT (Internet of Things). DataTrained. Retrieved
October 18, 2023, from https://datatrained.com/post/characteristics-of-iot/
Yampolskiy, R. V. (2016). Artificial SuperIntelligence: A futuristic approach. CRC Press.
Yudkowsky, E. (2006). Artificial intelligence as a positive and negative factor in global risk. In
́
N. Bostrom & M. M. Cirković (Eds.), Global catastrophic risks (pp. 308–345). Oxford University Press.
Zaveri, P. (2023, February 3). Microsoft’s landmark deal with OpenAI shows that ChatGPT is going to
be the defining technology of 2023. Business Insider. Retrieved October 19, 2023, from https://www.
businessinsider.com/why-generative-ai-chatgpt-the-defining-tech-of-this-year-2023-1?r=US&IR=T