Content uploaded by Ahmad Alsaggaf
Author content
All content in this area was uploaded by Ahmad Alsaggaf on Jan 30, 2025
Content may be subject to copyright.
Journal of Architecture and Planning
, Vol. 36 (3) , pp. 359-379, Riyadh ( 2024/1446H )
doi:10.33948/JAP-KSU-36-3-4
359
Generative Artificial Intelligence as a Source of Inspiration for
a Modern and Authentic Architectural Identity of Residential
Building Facades in Makkah City
Ahmad Ameen AlSaggaf1 Wadia Ali Albarqawi2 Sahl Abdullah Waheeb3
Saeed Ali Albarqawi1 Husam Murad2
1 Department of Urban and Engineering Research, the custodian of the two holy mosques Institute for Hajj and
Umrah Research, Umm Al-Qura University, Makkah, Saudi Arabia
2 Department of Islamic Architecture, Umm Al-Qura University, Makkah, Saudi Arabia
3 Applied College, Umm Al-Qura University, Makkah, Saudi Arabia.
aasaggaf@uqu.edu.sa, wabarqawi@uqu.edu.sa, sawahieb@uqu.edu.sa.,
saburgawi@uqu.edu.sa, hakmurad@uqu.edu.sa
(Received 15/2/2024; accepted for publication 4/7/2024.)
Abstract: Makkah City during the 19th century had rich architectural vocabularies and characteristics.
Due to the impact of modernism and the demolition of heritage buildings for the Holy Mosque
expansion, most of those traditional design vocabularies and characteristics disappeared from the
architectural practice of many designers in Makkah. These factors, among many others, have caused a
discontinuity between the current architectural attributes of Makkah and its rich architectural history,
representing the research problem and incentive. Through a descriptive and experimental methodology,
this paper examines the potential of generative artificial intelligence (GAI) as a source of inspiration
to resolve the research problem. The findings indicate that GAI can inspire and help practitioners
create a modern yet authentic architectural identity for residential building facades in Makkah City.
The reached identity is expected to improve the quality of life and social sustainability in Makkah City.
Keywords: Makkah; architecture; identity, artificial intelligence.
1. Introduction
Different cultures worldwide inspired the
Makkah architectural vocabularies (MAV) and their
characteristics during the 19th. Those architectural
features stemmed from the needs of the society to
accommodate for the harshly hot climate in the
Hijaz region in which the City of Makkah is located
(Baik and Boehm, 2017; UNESCO, 2014).
Today, the architectural vocabularies and
elements known in the cities of the Hijaz region
during the 19th century mostly disappeared due
to moving towards modernism (SCTA, 2013),
particularly in Makkah, where traditional buildings
were demolished to make space for the expansion of
the Holy Mosque. Information about the architectural
vocabularies and design elements of Makkah City
during the 19th century can only be found in books,
journal papers, websites, social media, and photo
albums. This caused an apparent discontinuity in
the architectural vocabularies between the existing
built environment in Makkah and its architectural
360 Ahmad AlSaggaf; Wadia Albarqawi; Sahl Waheeb; Saeed Ali Albarqawi; Husam Murad: Generative Artificial Intelligence...
heritage. As a result, the current architectural
identity of Makkah City lacks the authenticity
and grace of its rich and distinctive architectural
heritage. Additionally, the Saudi government, as
part of its national transformational plan called
“Vision 2030”, is encouraging the implementation
of architectural styles that combine modernity with
authenticity. To emphasize the country’s national
and regional characteristics, the Saudi government
introduced new building codes (RRM, 2022) and
established commissions to achieve this objective
(RCRC, 2022; RCU, 2022). The Governor of the
Mecca Region, HRH Prince Khaled Al-Faisal, has
directed adherence to the Meccan identity in urban
projects (SPA, 2016). Also, a working group has
been formed to enhance the architectural identity of
the Holy Capital (Makkah) (AlSalami, 2016), and
a draft working paper has been prepared to protect
the architectural identity of Makkah (MRDA,
2010). The CEO of the Royal Commission for the
Holy City of Makkah and The Holy Sites has also
announced the development of new neighborhoods
based on a distinctive architectural identity (SPA,
2023). In addition, researchers have suggested that
architects should reuse the 19th-century Hijazi
architectural vocabulary in their designs to sustain
and maintain continuity with the past (AL-BAN,
2016; Almaimani and Nawari, 2017; Almaimani
and Nawari, 2015a; Almaimani and Nawari 2015b;
Baik and Boehm, 2017). This is, however, currently
limited to only digitally replicating MAV and using
them in design via technology, which makes them
manifest as out of the context of modern design.
This paper examines the potential of utilizing
GAI as a new methodology to inspire and help
architects and designers create a contemporary
architectural identity that remains true to the
city’s authentic roots. It is worth mentioning that
architectural identity is a complex issue intertwined
between multiple elements; therefore, the focus
of this research will be limited to the facades of
19th-century residential buildings in Makkah City
and the most distinct architectural and design
vocabularies that characterize them as it is one of
the most crucial aspects of the local identity to be
achieved.
2. Literature review
2.1 Makkah’s architectural vocabularies during
the 19th century (MAV)
Makkah is a major city in the western region
of the Kingdom of Saudi Arabia. During the 19th
century, Makkah had unique and distinguishable
architectural elements and design vocabularies that
were in line with the values of the local society,
accommodated their requirements, and were the
results of a combination of factors, including social
(e.g., privacy), economical (e.g., using natural
Figure (1). An example of a traditional Makkan house from the 19th century (aoa3060). Modified by authors and students
361
Journal of Architecture and Planning
, Vol. 36 (3) , Riyadh (2024/1446H)
materials), and environmental (i.e., responding to
the harsh weather) (AL-BAN, 2016; Albarqawi,
2022; Al-Murahhem, 2010; Al-Murahhem 2011;
Al-Murahhem, 2022; Al-Murahhem, 2020; Al-
Shareef, 1996; Bagasi et al., 2021; Maghrabi, 2000;
SCTA, 2013; UNESCO, 2014). Those architectural
elements and design vocabularies were sustainable
solutions that sufficed the different needs of
local society and presented a unique architectural
identity, as will be discussed later in this paper.
After the discovery of oil in Saudi Arabia,
many factors, including the economic boom, the
move towards modernism, and the rapid urban
development, led to the disappearance of those
architectural elements and design vocabularies
in the present built environment in Makkah (Al-
Murahhem, 2010; Hamidaddin et al., 2020; SCTA,
2013; UNESCO, 2014) and consequently, is no
longer used in current architecture practice. Finding
information about the architectural features and
design language of 19th-century Makkah City can
be challenging. This information is typically found
in books, journal papers, websites, social media,
and photo albums. The most notable architectural
elements and design vocabularies that characterized
the facades of the 19th-century houses in Makkah
include Shorofat, Shawabeer, Roshan, and entrance,
as demonstrated in Figure 1.
This section provides a brief explanation
of each of the aforementioned architectural
characteristics and design language.
2.1.1 Shorofat
Shorofa (singular) and Shorofat (plural), also
known as Araies Alsamaa, are moldings usually set
on the top of the walls and parapet in an array and
at specific intervals for decoration purposes. They
come in various shapes and sizes and are commonly
colored white (Al-Murahhem, 2022)Sometimes,
they match the color of the wall or parapet they are
on. Figure 2 shows Shorofat in Alsaggaf Palace in
Makkah. Shorofa determined the skyline of many
traditional houses in Makkah.
Figure (2). Shorofat on top walls and parapets at Alsaggaf palace in Makkah (by student Hatem Mahjoob)
362 Ahmad AlSaggaf; Wadia Albarqawi; Sahl Waheeb; Saeed Ali Albarqawi; Husam Murad: Generative Artificial Intelligence...
2.1.2 Shawabeer
According to Hariri (1986), the rooftops of
Makkan residential buildings, the Kharja, served
as a nighttime extension of the dwellings. The
walls of the Kharja were built using Shawabir, who
was taller than a standing man. The Shawabeer
(plural), or Shabora (singular), is a treatment for
parapets or walls of Kharja (terrace open to the sky
typically located on upper floors). It is made of
bricks that are typically painted in various colors,
or sometimes they match the color of the wall or the
parapet they are located on (Al-Murahhem, 2022;
Al-Murahhem, 2020). The bricks are arranged
precisely to give an aesthetically pleasing look,
allow airflow through the small openings between
the stone bricks, facilitate sight from the inside, and
provide inhabitants with seclusion (Hariri, 1986),
as shown in Figure 3.
2.1.3 Roshan
Roshan is the most recognizable architectural
element in the Hejazi and Makkan buildings. The
Roushan is a wooden structure that covers the
apertures of residential houses in Makkah, according
to Angawi (1988). Alitany et al., (2013) describe it
as a large projecting three-sided wooden structure
on a facade with a recognizable latticed component.
It functions as a window and provides a seating and
sleeping area sufficient for adults (Al-Murahhem,
2010). Also, it protects occupants against harsh
environmental factors such as brightness from the
sun and the hot weather outside. As a social need, it
provides privacy for occupants as the Roushan has
little wooden screens called Qalalib that enable a
high level of seclusion because the outside cannot
view the internal rooms (Angawi, 1988). The colors
of the Roshan in Makkah typically include brown
and dark brown, depending on the wood it is made
of, as shown in Figure 4.
2.1.4 Entrance
The entrance of traditional houses in Makkah
was one of the notable architectural elements that
exhibited the unique Makkah architectural design
vocabulary from the 19th century. Typically, the
main entrance was decorated and recessed in the
house’s front wall, and an ornamented wooden door
was positioned in its center, as shown in Figures
5 and 6. Figure 6 shows the recessed entrance
providing enough waiting for guests.
As previously discussed, the MAV in the
Figure (3). Shawabeer at Alsaggaf Palace in Makkah (by student Hatem Mahjoob)
363
Journal of Architecture and Planning
, Vol. 36 (3) , Riyadh (2024/1446H)
residential building facades of 19th-century
Makkah City were sustainable solutions that met
various local needs while showcasing a distinctive
architectural style and identity. However, it should
be noted that, with the advent of technology in the
21st century, some of its functions, such as cooling
and ventilation, have become less important.
In contrast, other functions, such as providing
occupants privacy and protection from the harsh
weather, are still essential requirements for society
in Makkah City today, and having MAV-inspired
residential building façade design elements
Figure (4). Roshan at Alsaggaf palace in Makkah (by student Hatem Mahjoob)
Figure (5). Entrance at Alsaggaf Palace in Makkah (by student Hatem Mahjoob)
364 Ahmad AlSaggaf; Wadia Albarqawi; Sahl Waheeb; Saeed Ali Albarqawi; Husam Murad: Generative Artificial Intelligence...
and treatments can satisfy those requirements.
Moreover, in terms of the aesthetics of architectural
style, designing MAV-inspired residential building
facades will provide more cultural value than the
current western-inspired ones, as it will increase
the connection between contemporary design
solutions and the city’s architectural history, giving
historical depth and continuity in façades design in
Makkah City, which improves social sustainability
and complies with Saudi Arabia’s vision 2030.
2.2 Artificial Intelligence (AI) in Architecture
Artificial intelligence (AI) and using
computer models to solve problems in the field of
the Architectural, Engineering, and construction
(AEC) industry has been a topic of research since
the 60s (Tommelein et al., 1989). Furthermore,
practitioners, from both academia and the industry,
in the past thirty years, have been attracted to AI
techniques and computerized tools to solve problems
related to their field (Alsaggaf, 2020; Andayesh
and Sadeghpour, 2013). In the early studies, AI
models used what is known as optimization or
improving techniques. Those AI optimization
models can be categorized into mathematical,
heuristic, and knowledge-based methods (Farmakis
and Chassiakos, 2018). The most common AI
techniques, however, used in the studies done on
AEC industry are mathematical techniques and
heuristics techniques (Alsaggaf, 2020; Alsaggaf
and Jrade, 2017; AlSaggaf and Jrade, 2021;
Elbeltagi et al., 2005; Elbeltagi and Hegazy, 2001;
Hawarneh et al., 2021; Sanad et al., 2008). Each
approach has its advantages and disadvantages.
For instance, mathematical methods such as;
integer programming models, linear programming
models, mixed-integer linear programming, mixed-
integer non-linear programming, and dynamic
programming (Elbeltagi et al., 2005; Hawarneh et
al., 2021; Xu et al., 2021), models can generate
optimal solutions more accurately than heuristics
techniques; nevertheless, they have mathematical
limitations and take a long time to reach a solution,
and thus, they are better applied to small, well-
defined problems with fewer variables, while
heuristic methods such as; genetic algorithms,
ant colony optimization, bee colony algorithms,
and particle swarm optimization (Alsaggaf, 2020;
Alsaggaf and Jrade, 2017; AlSaggaf and Jrade,
2021) yield semi-optimal solutions; however, they
can deal with larger and more complex problems
faster than mathematical techniques. This advantage
of heuristic methods encouraged practitioners to
conduct extensive research on how to apply them
to solve different issues that have been facing the
AEC industry over the last two decades (Farmakis
and Chassiakos, 2018). Castro Pena et al., (2021)
conducted a study about the application of AI in the
field of architecture and reported an increase in the
number of publications by 85% from 2015 to 2020
alone. Most studies used optimization techniques,
including mathematical (i.e., cellular automata) and
heuristic (i.e., artificial neural networks, fractals,
swarm intelligence, and genetic algorithms). In
recent years, more than 73% of the studies used
heuristic techniques over mathematical techniques
to solve issues related to the form in architecture
at the conceptual phase, with genetic algorithms
accounting for almost 90% of the studies on the
topic and artificial neural networks being the least
investigated. This is because heuristic techniques
can include many variables and balance between
competing multiple criteria and objectives, such
as shape, cost, lighting, efficiency, sustainability,
acoustics, etc. while being faster at providing
a solution when compared with mathematical
techniques. Furthermore, the most investigated areas
in the literature on AI application in architecture
are; building shape, layout design, morphogenesis,
design exploration, floor plans, ceiling form, and
Figure (6). Entrance at Alsaggaf Palace in Makkah
(Makkawi, 2018)
365
Journal of Architecture and Planning
, Vol. 36 (3) , Riyadh (2024/1446H)
façade design, respectively. In addition, recent
studies seek to enhance the designs at hand, specially
form optimization, however; early studies looked
at design exploration and morphogenesis to inspire
designers by creating unusual and unforeseen
shapes. Since AI optimization techniques vary
in performance, a comparison between them
when applied to each of the aforementioned areas
of investigation in architecture is required to
identify the advantages and disadvantages of each
technique. Such a study is done, however, in the
AEC industry to compare the five evolutionary-
based optimization algorithms including Genetic
algorithms (GA), memetic algorithms (MA),
particle swarm optimization (PSO), ant colony
algorithm (SFLA) (Elbeltagi et al., 2005).
In the last two years, particularly since 2021,
new types of AI-based models known as “generative
artificial intelligence” models, also referred to as
“art AI,” have been emerging and have been heavily
used by practitioners in the architecture field from
academia and industry. The new generative AI
(GAI) models were pre-trained, using deep learning
methods, to understand the natural language from
textual input queries known as prompts and create,
as an output, new original images that satisfy the
requirements of these prompts (Ploennigs and
Berger, 2022). The said models are thus called text-
to-image (txt2img) models, which means they can
produce realistic artwork and graphics from natural
human language descriptions using prompts (Open
AI, 2022). A prompt serves as a framework for
creating images and carries within it the objectives
intended by users. For instance, in architecture,
users can ask the model using prompts to produce a
picture of a specific type of building with particular
building materials in a precise location, daytime, and
the style of a famous architect, and the model will
generate an image that complies with their requests.
The resulting illustrations of the GAI text-to-image
models are typically photorealistic 3D images with
sometimes such a high level of accuracy and detail
that makes them indistinguishable from actual
photographs (Nichol et al., 2021). Moreover, these
generative AI models can create images in a timely
fashion, typically in approximately 60 seconds,
which is another massive advantage of these
models. The practicality of the new generative AI
txt2img models, in addition to their robustness,
efficiency, accuracy, ease of use, and high-fidelity
results, all in a speedy manner, have made them
widely accepted since they were made available to
the public two years ago with the number of users
reaching millions from all over the world within the
last year and still increasing.
A very renowned example of the new
generative text-to-image AI models is Midjourney.
Since its release in March 2022 as a limited beta
(Midjourney, 2022a) and as an open beta in July
2022 (Midjourney, 2022b) through a server that can
be added as an extension to the Discord chatting
platform, Midjourney has gained global attention
from users, including architects and designers, who
have reached almost 14 million users (Discord,
2023) by the end of July 2023 (in 15 months
since its launch). The intuitive user interface of
the Discord messaging platform is used to write
the prompts to the Midjourney bot and share and
discuss the resulting artwork publicly and privately.
In an investigation of how Midjourney is used in
the architecture and design context, Ploennigs
and Berger (2022) analyzed 40 million queries of
the Midjourney users, 2.6 million of which were
architecture-related. In their analysis, they filtered
the queries based on a keyword list they had made
from the fields of architecture and interior design
and found; that the most and third most recurring
keywords were “architecture” and “interior”,
respectively. Among the least frequently quired
keywords was “façade”, ranking 15 out of 27.
Also, famous architects were repeatedly queried,
with Zaha Hadid being, by far, the most frequently
queried. Midjourney has shown the capability
of deducing users’ intentions with high accuracy
even if it was not prescribed as part of the prompt
written by users; therefore, it is highly beneficial
as a support tool to practitioners from academia
and industry in the early stages of the architectural
design process (Jaruga-Rozdolska, 2023). In
addition, GAI txt2img models can generate many
ideas and variations in a timely fashion, which
supports users during the ideation process. It is
still the responsibility of users, however, to provide
textual inputs that are easy to understand by the
generative AI tools to get the result intended by
them (Enjellina et al., 2023).
It is worth mentioning that, despite the
numerous advantages the GAI txt2img models
have, they still suffer from a few disadvantages
(Radhakrishnan, 2023). First, the capability of
GAI models to comprehend abstract ideas such as
beauty, privacy, space, and time and generate images
accordingly is questionable. Architects are aware of
366 Ahmad AlSaggaf; Wadia Albarqawi; Sahl Waheeb; Saeed Ali Albarqawi; Husam Murad: Generative Artificial Intelligence...
the sublime, which is a term for grand concepts like
life, death, the divine, and the spiritual dimension
(Nesbitt, 1995), and the great feeling they get in
response to being in a specific space and a particular
time and typically convey that while trying to find
a balance between aesthetics and function during
the design process. Second, the interpretation of the
sublime’s intricacies varies depending on the region
and culture. Regardless, GAI may not accurately
convey cultural significance in the generated
architectural illustrations. Finally, AI relies on huge
databases consisting of millions, even billions,
of images, but its capabilities are limited to the
content of those images. Accordingly, AI-generated
art demonstrates human actions, design, and
intervention using existing references. This means
that if these references contain discriminatory
patterns, the AI may unintentionally replicate them
in visual art. Also, GAI can yield aesthetically
pleasing architectural output but can be inaccurate
technical-wise (Jaruga-Rozdolska, 2023).
The literature shows that the building façade
and architectural identity have received the least
attention in published works on AI and architecture,
including those discussing GAI txt2img models.
Furthermore, while Generative AI models are
advanced in their ability to create, they still
struggle to comprehend the sublime or abstract
concepts and experience emotions associated with
a specific location or time, which humans excel at.
Also, producing culturally relevant images can be
challenging for generative artificial intelligence
models. These models are limited by the available
resources in their databases, which can affect the
genuineness of the images they produce and their
technical accuracy. In addition, as previously
stated, discovering details about the architecture
and design style of Makkah City in the 19th century
can be difficult, as this information is typically
located in books, journal articles, websites, social
media platforms, and photo albums. As a result,
generating a modern and authentic architectural
identity for Makkah using GAI technology poses a
challenging task.
3. Methodology
The current state of Generative Artificial
Intelligence (GAI) does not understand the logical
relationship between architectural objects. This
study aims to investigate the use of GAI in the initial
design phase of the building façade to create a blend
of modern and traditional architectural styles for
Makkah City. As the focus is on the ideation phase
of residential building façade design, it is assumed
that the functional requirements of the occupants
and space design have already been considered at
this stage. The paper aims to explore the potential of
GAI and its ability to inspire and assist architects in
creating the aforementioned architectural identity
for Makkah, with a particular focus on residential
building facades. We accomplished our objective
by taking the subsequent measures:
Figure (7). The process flow for generating the images using the Midjourney bot
367
Journal of Architecture and Planning
, Vol. 36 (3) , Riyadh (2024/1446H)
Identifying the most discriminative
architectural features and design languages that
characterized the facades of residential houses
in Makkah City during the 19th century (MAV)
by a) conducting a thorough review of literature,
journals, and books, b) analyzing photographs of
Makkah City before the impact of Modernism on
its architectural identity, using various sources
including websites and social media, and c) site
visits, we visited with our students Alsaggaf Palace
– the most renowned and only remaining traditional
residential building in Makkah that resembles, in
good condition, the said identified architectural
vocabularies. The students took photos of the
identified architectural elements.
This paper identified architectural elements
such as Shorofat, Shawabeer, Roshan, and the
entrance. We also pointed out their typical functions,
colors, materials, and shapes.
Using Midjourney as a GAI tool to generate
the desired modern and authentic architectural
identity for Makkah City. It is argued that it can
accurately determine user intentions, even if it was
not explicitly stated in the prompt. This makes it a
beneficial support tool for practitioners in academia
and industry during the initial stages of architectural
design. The prompts written to generate the images
and variations are a) A modern house in Makkah
with local architectural style, architectural elements,
and traditional materials from the 19th century at
daytime photorealistic, b) the same prompt was
used, but we removed “traditional materials” and
added “more glass” in the prompt, and, c) The same
previous two prompts were used, but we changed
the time in each prompt to “nighttime”. The authors
wanted to explore the capability of the GAI tool
to understand the different daytime and nighttime
prompts for variation purposes.
Both prompts were preceded by a predefined
operational prompt, /imagine, which is part of
Midjourney’s system. Figure 7 shows the process
flow for generating the images using the Midjourney
bot.
We intentionally did not mention in the
prompts any name of any of the architectural
elements and design vocabularies of Makkah
during the 19th century identified in this paper. The
reason is to see if the Midjourney bot can generate
the said elements in the design of a modern house in
Makkah without explicitly telling it to.
Also, we would like to test the ability of the
GAI model to provide genuine creative illustrations
that demonstrate a modern design while resembling
the traditional identity of 19th-century residential
building facades in Makkah City.
Examining the images generated by the
Midjourney bot. The analysis aimed to reveal
any resemblances or discrepancies between the
illustrations generated by the Midjourney bot and
the traditional architectural vocabularies of Makkah
identified in this paper while including them in a
creative modern design.
Then, based on the results of the analysis, we
reach one of the following verdicts;
GAI can completely replace architects and
designers in creating the desired authentic
and modern identity for Makkah City,
GAI can be a beneficial support tool for
users, to inspire and help them create a new
architectural identity for Makkah City that
stays true to its roots, and
GAI does not benefit practitioners as it lacks
the knowledge and capabilities required for
such a task.
4. Results
Figure 8 shows an image of modern houses
generated by the Midjourney bot based on prompt
1, described in the methodology. Figure 8 (a)
demonstrates some resemblance to MAV (the
architectural components and design vocabularies
from 19th-century Makkah City identified in this
paper as described in the methodology), namely
Shawabeer (in a Green box) and the decorated
recessed entrance. Figures 8 (b) and (c) all show
similarities to MAV with a creative representation
of the Roshan on the windows, especially in Figure
8 (c), where the entrance and the window and
balcony on top of it are all beautifully connected
with the Roshan-inspired wooden structure. Also,
both Figures show gorgeously decorated recessed
entrances. Figure 8 (d) shows the most resemblance
to MAV compared to all the previous Figures.
A clear creative illustration of the Roshan-like
structure connects the upper windows with the
bottom ones and the upper window and balcony
with the recessed ornate entrance. Also, we can
see a Shawabeer-inspired element in the parapet in
addition to Shorofat-inspired moldings on top of
the parapet.
Prompt 1 was also used to generate Figure
9. Figures 9 (a) and (b) show elegant and creative
illustrations for Shawabeer, Roshan, and entrance,
368 Ahmad AlSaggaf; Wadia Albarqawi; Sahl Waheeb; Saeed Ali Albarqawi; Husam Murad: Generative Artificial Intelligence...
Figure (8). An image generated by the Midjourney bot based on prompt 1, described in the methodology
Figure (9). An image generated by Midjourney based on prompt 1, described in the methodology
369
Journal of Architecture and Planning
, Vol. 36 (3) , Riyadh (2024/1446H)
Figure (10). An image generated by the Midjourney bot based on prompt 1, described in the methodology
Figure (11). A variation of image 8 (c) generated by the Midjourney bot
370 Ahmad AlSaggaf; Wadia Albarqawi; Sahl Waheeb; Saeed Ali Albarqawi; Husam Murad: Generative Artificial Intelligence...
nevertheless, there are no elements resembling
Shorofat. Figure 9 (c) depicts what can be seen as
a creative modern take on Roshan. The room above
the entrance is extruded compared to the other rooms
on the same level, and the latticed metal cover is
placed in front of the glass windows resembling
the same design and functions of Roshan. The
Figure illustrates a recessed ornamented entrance
and recessed windows surrounded by blue extruded
walls with stone ornamentation, all with vector-
shaped arch designs on their top. Also, two
components on top of the house, one on each front
corner, demonstrate a design inspired by Shorofat.
In Figure 9 (d), we can see some Roshan-inspired
decorations covering the balcony’s fence and the
bottom parts of the windows. In addition, elements
that resemble Shorofat on the parapet’s top seem to
be inspired by the minarets of the mosque. Other
than that, no relation to MAV is noticed.
Figure 10 shows another iteration of
using prompt 1. We can see in Figure 10 (a) the
illustration of a beautiful Roshan-like component
that connects the upper balcony with the main
entrance. Also, in the parapet, we see an element
(in a Green box) inspired by the Shawabeer. The
design of the windows is also beautifully in line
with the layout of the Roshan. In Figure 10 (b), the
design of the windows and main entrance gracefully
resembles the Roshan layout and embellishments;
however, some strange components in the red box
do not match the overall design of the house and
seem out of place. In Figures 10 (b) and (c), the
entrance is recessed with adornment and an arch
design. Especially in Figure 10 (c), it is stunningly
decorated, and the way it is connected to the above
glass window resembles Roshan in addition to the
top windows to the right and left of the entrance
below, where this resemblance is apparent. Also,
a Shawabeer-like element, in the green box, with
a creative design resides in the parapet on the left
side of the house. Shorofat-inspired moldings in
Figures 10 (c) and (d) tastefully decorate the top of
the parapet. In Figure 10 (d), the recessed decorated
entrance is not displayed; however, the arched door
design matches that of MAV. With that said, Figure
10 (d) does not illustrate a resemblance to MAV.
With Midjourney’s bot, users have the option
to create four different variants of a selected image
by using the variant prompt. They can also choose
between strong or subtle variations to display in
the final result. Figures 11 and 12 represent strong
variations of Figures 8 (c) and 9 (c), respectively,
Figure (12). A variation of image 9 (c) generated by the Midjourney bot
371
Journal of Architecture and Planning
, Vol. 36 (3) , Riyadh (2024/1446H)
Figure (13). A strong variation of image 10 (c) generated by the Midjourney bot
Figure (14). A subtle variation of image 10 (c) generated by the Midjourney bot
372 Ahmad AlSaggaf; Wadia Albarqawi; Sahl Waheeb; Saeed Ali Albarqawi; Husam Murad: Generative Artificial Intelligence...
Figure (15). An image generated by the Midjourney bot based on prompts 2 and 3, described in the methodology
Figure (16). An image generated by the Midjourney bot based on prompts 2 and 3, described in the methodology
373
Journal of Architecture and Planning
, Vol. 36 (3) , Riyadh (2024/1446H)
while Figures 13 and 14 are strong and subtle
variations of Figure 10 (c), respectively. The
strong variations, represented in Figures 11, 12,
and 13, showcase all the similarities to MAVs in
the original images while also featuring unique and
radical changes to their layout, forms, locations,
ornamentation, and overall design of the house.
On the other hand, Figure 14 represents a subtle
variant, where the house layout in the variants
is kept similar to the original image, particularly
the locations of doors and windows. The features
and elements that resemble Shorofat, Shawabeer,
Roshan, and the entrance, were creatively and
artistically redesigned in terms of their forms and
adornments.
During nighttime, Figures 15 and 16 exhibit
a modern design style with a predominant use
of glass on the facade, as opposed to traditional
materials. The “more glass” prompt replaced the
earlier prompt for “traditional materials”, leading
to a reduction in the use of wood and stone in the
resulting images. Although the designs bear less
resemblance to MAV, they still feature some design
vocabularies and functions. For example, Figure
15 showcases an arch design for windows and
doors, while the glass facades in Figures 15 and
16 are adorned with beautiful latticed components
that offer privacy and protection against heat and
sunlight glare, similar to the functions of Roshan.
5. Discussion
Creating a new yet authentic identity
for Makkah City using Generative Artificial
Intelligence (GAI) faces four main challenges.
Firstly, GAI struggles to understand, as humans
do, abstract concepts and emotions associated with
specific locations or times. Secondly, producing
culturally relevant images can be difficult for GAI.
Thirdly, the limited resources in their databases can
affect the authenticity and technical accuracy of the
result images. Lastly, finding information about the
architecture and design style of Makkah City in the
19th century can be challenging as it is often found
in books, articles, websites, social media platforms,
and photo albums. These challenges make it difficult
to generate a modern and authentic architectural
identity for Makkah using GAI technology.
Thus, in this paper, we examined the
capabilities of GAI in generating useful images
that illustrate the desired identity. Our tests show
that the GAI tool (Midjourney’s bot) succeeded
in producing useful illustrations that demonstrate
a creative integration between modern design and
traditional 19th-century Makkah architectural
vocabularies (MAV). Our tests also revealed that
the GAI tool was capable of accomplishing this
task without explicitly naming any of the MAVs
(e.g., Shorofat, Shawabeer, Roshan, the entrance)
or the traditional materials (e.g., wood or stone). In
addition, Figures 9, 10, 11, and 12 show how the GAI
tool can create elements related to the environment
of Makkah, such as the surrounding mountains.
Also, the house shown in Figure 17 has blue walls
(and the one illustrated in Figure 9 C), similar to
the shade of blue that was sometimes used to paint
traditional homes in Makkah, typically painted in
beige or white colors, during the 19th century. This
suggests the GAI tool can deduce users’ intentions
and requirements from implicit texts. It remains
the responsibility of users to write what they are
looking for correctly.
We noticed that almost all generated images
demonstrated some manifestation of Roshan-
like design elements in creative and inspiring
illustrations. The least resembled MAV in the
generated pictures were Shorofat and Shawabeer.
This can be because Roshan is the most dominant
architectural element in 19th-century Makkah City
compared to other MAVs. It is worth mentioning
that all the strong and subtle variants maintained
the same design essence as the original images,
including the resemblance to MAV, colors, and
materials. Users can select the image that best
represents their desired outcome and produce
variants that creatively offer more diverse
illustrations of their desired features.
The GAI tool can generate illustrations based
on textual prompts, but it occasionally produces
strange features, as seen in Figure 18’s anomalies
(highlighted in red boxes). These anomalies, such as
the cut and floating stairs (Figure 18 a) or the stairs
that lead to a solid floor (Figure 18 b), showcase
GAI’s inability to understand the technical and
logical relationship between architectural objects
in its current stage.
After examining the GAI tool, we select
verdict B (explained in the methodology section) as
we found that, in its current state, it is a beneficial
support tool for users from industry (e.g., architects
and designers) and academia (i.e., students and
educators) that will help and inspire them to
create the desired identity for Makkah City that
is explained in this paper. We also note that users
374 Ahmad AlSaggaf; Wadia Albarqawi; Sahl Waheeb; Saeed Ali Albarqawi; Husam Murad: Generative Artificial Intelligence...
must know the architectural vocabulary and design
elements of Makkah City during the 19th century
(MAV). This helps them select the best illustrations
and variants resembling MAV.
Conclusions:
In the 19th century, Makkah City boasted
unique architectural vocabularies (MAV). As
modernism and other factors have taken hold, these
MAVs have fallen out of use by current architectural
designers. Information regarding MAV is solely
available in books, journal papers, social media
websites, and photo albums. This has resulted in
a break between Makkah’s current architectural
characteristics and its rich history. To address this,
a new and authentic architectural identity is needed
to improve the social sustainability and quality
Figure (17). An image generated by the Midjourney bot showcases walls with a blue color similar to what was used on the walls
of traditional houses in Makkah City during the 19th century.
Figure (18). Examples of anomalies in the images generated by the Midjourney bot
375
Journal of Architecture and Planning
, Vol. 36 (3) , Riyadh (2024/1446H)
of life in Makkah City and honor its traditional
architectural heritage. GAI offers potential for
creating this identity, nevertheless, it is evident
from the literature that architectural identity and
building façade were the least examined topics. This
can, in part, be tied to the inherited shortcomings of
the earlier AI models, also known as optimization
models. The first quarter of 2022 witnessed the
emergence of a new generation of AI models called
Generative Artificial intelligence (GAI). The GAI
is a promising technology that can tackle topics
such as architectural identity and building façade
design, yet there are a few challenges that hinder
the ability of GAI to offer an acceptable solution to
represent an architectural identity, especially in the
case of Makkah City. This paper investigates using
GAI as a novel methodology to inspire and assist
architects and designers while creating a modern
and authentic architectural identity for Makkah
City. The findings suggest that GAI can serve as an
inspiration source and help architects and designers
create the desired architectural identity, which
will enhance social sustainability and improve the
quality of life in the city.
6. References
Arabic references:
Al-Murahhem, Faredah. “Kharja in the traditional
house - the solution to the Corona crisis”.
Journal of Architecture and Planning - King
Saud University, 34(1). (2022).
English references:
Abbas, Hidaya. “Catalogue of Hijazi Architecture
in Jidda and Makkah c . 1850-1920s :
Selected Buildings” (Issue May). The
American University in Cairo. (2014).
AL-BAN, Alaa. Z. G. “ARCHITECTURE
AND CULTURAL IDENTITY IN THE
TRADITIONAL HOMES OF JEDDAH”
[University of Colorado]. (2016).
Albarqawi, Wadia. “The State of Becoming of
the Roshan: Villa Aladdin and the Magic
Courtyard”. American Journal of Civil
Engineering and Architecture, 10(2),
101–105. (2022). https://doi.org/10.12691/
ajcea-10-2-5
Alitany, Ayman., Redondo, E., & Adas, A. “THE
3D DOCUMENTATION OF PROJECTED
WOODEN WINDOWS (THE ROSHANS)
IN THE OLD CITY OF JEDDAH (SAUDI
ARABIA) USING IMAGE-BASED
TECHNIQUES”. ISPRS Annals of the
Photogrammetry, Remote Sensing and
Spatial Information Sciences, Volume
II-5/W1, 2013 XXIV International CIPA
Symposium. (2013, September 2). https://
doi.org/https://doi.org/10.5194/isprsannals-
II-5-W1-7-2013
Almaimani, Ayad. K., & Nawari, N. O. “BIM-
Driven components library for Islamic
Facilities (BIM-IF)”. Visualization in
Engineering, 5(1). (2017). https://doi.
org/10.1186/s40327-017-0041-6
Almaimani, Ayad. M., & Nawari, A. N. O. “BIM-
driven islamic construction: Part 2 - Digital
libraries”. Congress on Computing in Civil
Engineering, Proceedings, 2015-January
(January), 460–467. (2015a). https://doi.
org/10.1061/9780784479247.057
Almaimani, Ayad. M., & Nawari, N. O. “BIM-
driven islamic construction: Part 1 - Digital
classification”. Congress on Computing
in Civil Engineering, Proceedings,
2015-Janua(January), 453–459. (2015b).
https://doi.org/10.1061/9780784479247.056
Al-Murahhem, Faredah. “The mechanism of
WIT Transactions on Ecology and the
Environment, 128, 561–573. (2010). https://
doi.org/10.2495/ARC100481
Al-Murahhem, Faredah.
Ornamentation as a Source of Creativity”.
Design Principles and Practices: An
International Journal, 5(6). (2011). http://
www.Design-Journal.com, https://doi.
org/10.33948/jap-ksu-34-1-2
Al-Murahhem, Faredah. M. “kharja” the lung of
the traditional house—the case of Makkah.”
In Advances in Science, Technology and
Innovation (pp. 45–52). (2020). Springer
Nature. https://doi.org/10.1007/978-3-030-
17308-1_4
376 Ahmad AlSaggaf; Wadia Albarqawi; Sahl Waheeb; Saeed Ali Albarqawi; Husam Murad: Generative Artificial Intelligence...
Alsaggaf, Ahmad. “ArcSPAT: An integrated
BIM-GIS model for site layout planning”
[University of Ottawa]. (2020). https://doi.
org/http://dx.doi.org/10.20381/ruor-24420
Alsaggaf, Ahmad., & Jrade, A. “A framework
for an integrated BIM-GIS decision support
model for site layout planning”. Proceedings
of Canadian Society for Civil Enginnering,
Leadership in Sustainable Infrastructure,
6th International Construction Specialty
Conference, 1–11. (2017).
AlSaggaf, Ahmad., & Jrade, A. “ArcSPAT: an
integrated building information modeling
(BIM) and geographic information system
(GIS) model for site layout planning.”
International Journal of Construction
Management, 0(0), 1–25. (2021). https://doi.
org/10.1080/15623599.2021.1894071
AlSalami, Faisal. “6
”. Makkahnewspaper.Com.
(2016). https://makkahnewspaper.com/
article/146801/6-
Al-Shareef, Faisal. M. “Natural light control in
Hedjazi architecture: an investigation of
the Rowshan performance by computer
simulation”. In PQDT - UK & Ireland.
(1996).
Andayesh, Mohsen., & Sadeghpour, F. “Dynamic
site layout planning through minimization
of total potential energy”. Automation in
Construction, 31, 92–102. (2013). https://
doi.org/10.1016/j.autcon.2012.11.039
Angawi, Sami. Mohsen. “Makkah Architecture”.
University of London. (1988).
aoa3060. “ ”.
Retrieved July 18, 2023, from https://
www.flickriver.com/photos/100992408@
N07/12545180555/#large
Bagasi, Abdullah. A., Calautit, J. K., & Karban,
A. S. “Evaluation of the integration of the
traditional architectural element mashrabiya
into the ventilation strategy for buildings
in hot climates”. Energies, 14(3). (2021).
https://doi.org/10.3390/en14030530
Baik, A., & Boehm, J. “Hijazi architectural object
library (HAOL)”. International Archives
of the Photogrammetry, Remote Sensing
and Spatial Information Sciences - ISPRS
Archives, 42(2W3), 55–62. (2017). https://
doi.org/10.5194/isprs-archives-XLII-
2-W3-55-2017
Castro Pena, Maria. L., Carballal, A.,
Rodríguez-Fernández, N., Santos, I., &
Romero, J. “Artificial intelligence applied
to conceptual design. A review of its use in
architecture”. In Automation in Construction
(Vol. 124). Elsevier B.V. (2021). https://doi.
org/10.1016/j.autcon.2021.103550
Discord. “Discord | Home – Midjourney”. Retrieved
July 31, 2023, from https://discord.com/
channels/662267976984297473/@home
Elbeltagi, Emad., & Hegazy, T. “A Hybrid AL-
Based System for Site Layout Planning in
Construction”. Computer-Aided Civil and
Infrastructure Engineering, 16(2), 79–93.
(2001). https://doi.org/10.1111/0885-
9507.00215
Elbeltagi, Emad., Hegazy, T., & Grierson, D.
“Comparison among five evolutionary-
based optimization algorithms”. Advanced
Engineering Informatics, 19(1), 43–
53. (2005). https://doi.org/10.1016/j.
aei.2005.01.004
Enjellina, Vilgia Putri Beyan, E., & Gisela
Cinintya Rossy, A. “A Review of AI Image
Generator: Influences, Challenges, and
Future Prospects for Architectural Field”.
JARINA - Journal of Artificial Intelligence
in Architecture, 2(1). (2023). ojs.uajy.ac.id/
index.php/JARINA
Farmakis, Panagiotis. M., & Chassiakos, A.
P. “Genetic algorithm optimization for
dynamic construction site layout planning”.
Organization, Technology and Management
in Construction: An International Journal,
10(1), 1655–1664. (2018). https://doi.
org/10.1515/otmcj-2016-0026
Hamidaddin, Abdulkareem., Barashed, K., &
Al salafi, J. “The Impact of Contemporary
Developments on the Makki House (For the
Extened Family)”. Journal of Umm Al-Qura
University for Engineering and Architecture
, 11(2), 18–29. (2020).
377
Journal of Architecture and Planning
, Vol. 36 (3) , Riyadh (2024/1446H)
Hariri, Majdi. M. “Housing in Central Makkah the
influence of Hajj” [Newcastle University].
(1986). http://hdl.handle.net/10443/912
Hawarneh, Alaa. A., Bendak, S., & Ghanim, F.
“Construction site layout planning problem:
past, present and future”. Expert Systems
With Applications, 168(114247). (2021).
https://doi.org/https://doi.org/10.1016/j.
eswa.2020.114247
Jaruga-Rozdolska, Anna. “Artificial intelligence
as part of future practices in the architect’s
work MidJourney generative tool as part of
a process of creating an architectural form”.
Architectus, 3(71), 95–104. (2023). https://
doi.org/10.37190/arc220310
Maghrabi, Amjed. “Airflow characteristics of
modulated louvered windows with reference
to the Rowshan of Jeddah, Saudi Arabia”
(Issue August). University of Sheffield.
(2000).
Makkawi.
”.
Retrieved August 4, 2023,
Midjourney. “Midjourney on Twitter: Signups
for our limited beta are open”. Retrieved
August 2, 2023, from https://twitter.com/
midjourney/status/1505993505537081345
Midjourney. “Midjourney on Twitter: Today we’re
starting to test our V3 image generation
algorithms”. Midjourney. Retrieved
August 2, 2023, from https://twitter.com/
midjourney/status/1551560891254001664
MRDA. “
2010 ). https://www.
geocities.ws/fotouh28/H-Aoulfotouh-
Makkah-local-Identity.pdf
Nesbitt, Kate. “The Sublime and Modern
Architecture: Unmasking (an Aesthetic
of) Abstraction”. 83RD ACSA ANNUAL
MEETING, 177–182. (1995).
Nichol, Alex., Dhariwal, P., Ramesh, A., Shyam,
P., Mishkin, P., McGrew, B., Sutskever, I.,
& Chen, M. “GLIDE: Towards Photorealistic
Image Generation and Editing with Text-
Guided Diffusion Models”. (2021). http://
arxiv.org/abs/2112.10741
Open AI. “DALL·E 2”. OpenAI. Retrieved July
31, 2023, from https://openai.com/dall-e-2
Ploennigs, Joern., & Berger, M. “AI Art in
Architecture”. (2022). http://arxiv.org/
abs/2212.09399
Radhakrishnan, Ar. Mohsen. “
”. GSJ, 11(1). (2023). www.
globalscientificjournal.com
RCRC. ””. Retrieved August 8, 2023, from
https://www.rcrc.gov.sa/WadiUrbanCode/
RCU. “ ”. Retrieved August 8,
2023, from https://uds.rcu.gov.sa/
RRM. “ ”. Retrieved
August 8, 2023, from https://eservices.
alriyadh.gov.sa/Pages/BLS/Common/
BuildingCodes.aspx
Sanad, Haytham. M., Ammar, M. a, & Ibrahim,
M. E. “Optimal Construction Site Layout
Considering Safety and Environmental
Aspects”. Journal of Construction
Engineering and Management, 134(7),
536–544. (2008). https://doi.org/10.1061/
(ASCE)0733-9364(2008)134:7(536)
SCTA. “Historic Jeddah, the Gate to Makkah”.
(2013).
SPA . “
”. General. https://www.spa.gov.
sa/1489571 (2016, April 14).
SPA . “
2023 , January 22).
General. https://www.spa.gov.sa/w1842794
Tommelein, Iris. D., Levitt, R. E., & Hayes-Roth,
B. “Sightplan: an Artificial Intelligence Tool
To Assist Construction Managers With Site
Layout”. on Automation and Robotics in
Construction, June 6-8, 340–347. (1989).
http://medcontent.metapress.com/index/
A65RM03P4874243N.pdf\nhttp://www.
iaarc.org/publications/fulltext/Sightplan_
an_artificial_intelligence_tool_to_assist_
construction_managers_with_site_layout.
PDF
378 Ahmad AlSaggaf; Wadia Albarqawi; Sahl Waheeb; Saeed Ali Albarqawi; Husam Murad: Generative Artificial Intelligence...
Unesco. “Historic Jeddah, the Gate to Makkah”.
(2014). https://whc.unesco.org/en/list/1361/
Xu, Maozeng., Mei, Z., Luo, S., & Tan, Y.
“Optimization algorithms for construction
site layout planning: a systematic literature
review”. Engineering, Construction and
Architectural Management, 27(8), 1913–
1938. (2021). https://doi.org/10.1108/
ECAM-08-2019-0457
379
Journal of Architecture and Planning
, Vol. 36 (3) , Riyadh (2024/1446H)
sawahieb@uqu.edu.sa, wabarqawi@uqu.edu.sa, aasaggaf@uqu.edu.sa,
hakmurad@uqu.edu.sa, saburgawi@uqu.edu.sa
////
GAIGAI
doi:10.33948/JAP-KSU-36-3-4
