Transformative integration of Artificial Intelligence in architectural design

Małgorzata Kurcjusz, Anna Stefańska

doi:10.37190/arc250109

Abstract

The introduction of artificial intelligence (AI) into architectural design is changing traditional practices. This article presents a typological classification of AI applications in architecture, focusing on their contribution to conceptual design, performance optimization, and sustainability enhancement, highlighting the transformative impact on processes in the architecture, engineering, and construction (AEC) sector. The authors’ aim was to classify and analyze hierarchical applications of AI in architectural design. The main purpose was to assess AI’s role in enhancing creativity, increasing efficiency, and supporting sustainable development practices. Methodologically, the study was based on a comprehensive review of typological classifications, experimental validations, and case study analyses. The synthesis of results from recent literature and actual implementations allows for the evaluation of AI’s potential and limitations at various stages of architectural design.

Full article view is only available on bigger screens.

2025

1(81)

Małgorzata Kurcjusz*, Anna Stefańska**

Transformative integration of Artificial Intelligence

in architectural design

DOI: 10.37190/arc250109

Published in open access. CC BY NC ND license

Abstract

The introduction of Articial Intelligence (AI) into architectural design is transforming traditional practices. This paper presents the typological

applications of AI in architecture, focusing on its contributions to conceptual design, performance optimization, and sustainability, highlighting its

transformative impact on workows in the Architecture, Engineering, and Construction (AEC) sector.

The aim of the authors of the paper was to categorize and analyse the hierarchical applications of AI in architectural design. The primary objective

– to evaluate AI’s roles in fostering creativity, enhancing eciency, and supporting sustainable practices. Methodologically, the research incorpo-

rates a comprehensive review of typological classications, experimental validations, and case study analyses. By synthesizing ndings from recent

literature and real-world implementations, the paper assesses both the potentials and limitations of AI across various stages of architectural design.

The results of the study indicate a typological hierarchy of AI applications in architectural design, revealing its multifaceted contributions. In

conceptual design, AI-driven generative tools empower architects to explore vast design spaces, enabling the creation of innovative and diverse

solutions. In performance optimization, AI demonstrates signicant eciency gains by reducing design iteration times, minimizing material waste,

and enhancing energy eciency through advanced simulations. For sustainability, AI’s predictive analytics facilitate informed decisions on material

selection and compliance with environmental standards, supporting the achievement of green building certications. However, these advancements

come with challenges, including a dependency on high-quality data, ethical concerns regarding decision accountability, and the demand for special-

ized expertise. Based on the results, it can be concluded that AI enhances creativity, precision, and sustainability in architecture, but its successful

integration requires a collaborative human-AI approach, supported by robust ethical frameworks and ongoing skill development.

Key words: Articial Intelligence (AI), algorithms, parametric equations, contemporary architectural design

Introduction

Architectural design is associated with the ingenuity and

creativity of the human mind. Traditional design process-

es have relied heavily on manual calculations, experience,

and intuition to create functional and aesthetically pleasing

structures. However, these traditional methods often need

to be revised regarding eciency, precision, and the ability

to explore various design alternatives needed in the con-

temporary multicriterial optimisation in the AEC (Archi-

tecture, Engineering, Construction) sector. In recent years,

Articial Intelligence (AI) has started to impact architec-

tural design, paving the way for signicant changes in the

industry. AI’s capacity for handling complex algorithms

and large datasets enables architects to explore innovative

possibilities, venture into unconventional design territo-

ries, and discover groundbreaking solutions. AI-driven

tools can simulate numerous design scenarios in a fraction

of the time it would take through traditional methods, pro-

viding architects with valuable insights and data-driven

recommendations to streamline their creative process.

To fully understand the role of AI in architectural de-

sign, it is essential to outline its levels of sophistication and

recognise its current stage of development. AI is generally

categorised into three stages: Articial Narrow Intelligence

(ANI), Articial General Intelligence (AGI), and Articial

Superintelligence (ASI). Presently, the architectural indus-

try operates within the ANI stage, where AI systems are

designed to perform specic tasks within a limited scope.

* ORCID: 0000-0002-9703-2118. Institute of Civil Engineering,

Warsaw University of Life Sciences, Poland.

** ORCID: 0000-0002-5070-3877. Institute of Civil Engineering,

Warsaw University of Life Sciences, Poland, e-mail: anna_stefanska@

sggw.edu.pl

90 Małgorzata Kurcjusz, Anna Stefańska

Examples include image recognition tools, recommenda-

tion systems, and simulation algorithms designed to opti-

mize energy eciency or material utilization. These tools,

classied as ANI, improve performance and aid deci-

sion-making, but are limited by a lack of creativity and an

inability to adapt beyond predened tasks. The next stage,

AGI, represents a level where AI can perform diverse

tasks at a human-like level, including solving complex de-

sign challenges autonomously. Researchers have proposed

classifying AGI into levels like competent, expert, and

superhuman, depending on the extent of its capabilities.

Behind AGI lies ASI, a theoretical concept that refers to AI

that surpasses human intelligence in all domains, includ-

ing creativity, decision-making, and problem-solving. At

this level, AI would possess unparalleled analytical capa-

bilities and the ability to self-improve, potentially revolu-

tionizing elds like architecture. Currently, AI applications

in architecture remain focused on ANI, where they assist

with generative design, energy simulations, and predictive

maintenance. Recognising these stages helps clarify AI’s

current limitations and its transformative potential for the

future of architectural design (Abioye et al. 2021).

However, AI’s role in architectural design is not limited

to the creative process as an illustration-generating engine.

It extends to various stages of project development, from

initial concept to nal construction and building opera-

tion systems. AI applications assist in optimising building

performance, enhancing energy efficiency, and ensur-

ing structural integrity. They also facilitate better project

management by predicting potential issues, streamlining

stakeholder communication, and automating routine tasks.

The architectural design is currently undergoing a signif-

icant transformation brought about by the emergence of

AI. While rich in history and creativity, traditional design

methodologies often fall short in eciency, adaptabili-

ty, and the capacity to explore many design alternatives

(Fig. 1). Integrating AI in architectural design heralds

a new epoch where these limitations are being addressed,

paving the way for groundbreaking innovations and solu-

tions (Kineber et al. 2024). The utilisation of AI extends

beyond the realm of ideation, permeating every facet of

project development, from conception to implementation.

Applications of AI in architectural design

Generative AI holds signicant promise for revolu-

tionising structural design in architecture. It uses machine

learning algorithms to analyse historical designs, assim-

ilate diverse data types, and craft new, optimised struc-

tural congurations. This technology can greatly improve

both eciency and innovation by automating the layout

process, optimising material use, and complying with me-

chanical and empirical standards, which in turn reduces de-

sign times and costs (Liao et al. 2024). Architecture meets

advanced technology, where AI is pivotal in oering inspi-

ration and innovative solutions. AI-based algorithms en-

able the creation of diverse and innovative design options,

broadening the horizons for designers. Furthermore, they

expedite the design process by allowing for quick analysis,

optimisation, and forecasting of project-related aspects.

AI signicantly aids in enhancing the energy eciency

of buildings through energy simulations and identifying

and eliminating design aws at early stages. By analysing

customer preferences, AI tailors design to meet individual

needs, increasing time eciency in the designing process.

The algorithms assist in cost analysis budget management

and improve collaboration with other elds, such as struc-

tural engineering and urban planning. AI also advances re-

search on innovative materials and structures, paving the

way for groundbreaking and sustainable solutions (Elwy,

Hagishima 2024). It supports decision-making processes

through data analysis and helps identify design trends.

The construction industry faces signicant challenges,

such as cost and time overruns, health and safety issues,

productivity issues, and labour shortages, limiting its

growth and productivity. One major issue is the lack of op-

timization; it is one of the least optimized industries glob-

ally, leading to ineciencies in project management, cost

control, and decision-making processes. AI technologies,

including machine learning, computer vision, and robot-

ics, have optimized other industries like manufacturing,

retail, and telecommunications by improving protability,

eciency, safety, and security. These technologies pres-

ent numerous opportunities for the construction industry,

including activity monitoring, risk management, resource

and waste optimisation, and enhancing overall project e-

ciency (Ganiyu et al. 2020). Various AI techniques are al-

ready being used in construction, such as machine learning

for health and safety monitoring, cost estimation, risk pre-

diction, and robotics for site monitoring and performance

evaluation.

Future AI applications could include AI-driven waste

analytics, AI-integrated Building Information Modeling

(BIM) for cost and time prediction, and AI-powered con-

Fig. 1. Major contributions of AI

to the architecture domain.

Image on the right was generated

by AI ChatGPT-4,

using the Image Generator

(elaborated by M. Kurcjusz)

Il. 1. Główny wkład AI

w dziedzinę architektury.

Obraz po prawej stronie

został wygenerowany przez AI

ChatGPT-4 przy użyciu Image

Generator (oprac. M. Kurcjusz)

 TransformativeintegrationofArticialIntelligenceinarchitecturaldesign 91

struction site analytics. BIM is a methodology and tool for

creating a comprehensive digital building model incor-

porating geometric, material, and functional information.

However, in its basic form, BIM relies on programming,

data management, and rule-based processes rather than AI.

While BIM and AI are distinct technologies, combining

them can signicantly enhance the eciency of building

design, construction, and management (Abdulfattah et al.

2023). Integrating AI algorithms – such as machine learn-

ing, data analysis, or generative design – into BIM work-

ows would make it possible to create more sophisticated

tools that enable smarter design, advanced analysis, and

more ecient data management. In such cases, AI acts

as a supporting element that augments the capabilities of

BIM, transforming it into a more intelligent and predictive

tool. Therefore, while BIM does not inherently qualify as

AI, it can benet from AI technologies to achieve greater

automation, optimization, and innovation throughout the

design and construction process (Khan et al. 2024).

However, key barriers to AI adoption in construction

include cultural resistance to change, high initial deploy-

ment costs, security concerns, and the need for explainable

AI. AI has the potential to signicantly improve health and

safety in construction through predictive analytics, inte-

gration with wearable technology, and advanced data an-

alytics to foresee and mitigate risks. Additionally, AI can

enhance construction contract management by automating

tasks like contract optimizing and dispute resolution and

ensuring compliance, thus reducing human error and in-

creasing eciency (Abioye et al. 2021).

AI algorithms are optimizing the eld of architectural

design by enabling the generation of diverse and inno-

vative design options, thus opening new perspectives for

designers. These algorithms facilitate the analysis, optimi-

zation, and prediction of various design-related aspects,

tailoring projects to individual needs. AI supports cost

analysis and budget management, enhancing collabora-

tion with other disciplines. Furthermore, AI aids research

into innovative materials and constructions, paving the

way for novel and sustainable solutions. It also supports

decision-making processes by providing data analyses

and identifying design trends. AI technologies are making

signicant contributions to architectural design. AI-based

tools assist architects in generating design options based

on specic criteria such as client requirements, site condi-

tions, and budget constraints. AI enhances user experience

and human-centred design by analysing user behaviour

and preferences, enabling more human-focused design

decisions. AI-driven building systems can adapt to chang-

ing environmental conditions and user needs, optimising

lighting and temperature based on occupancy and time of

day. Moreover, AI can analyse building performance data,

including energy consumption, daylighting, and thermal

comfort, helping architects optimize energy eciency and

sustainability designs. Architects can create more environ-

mentally friendly and cost-eective buildings by consider-

ing these factors early in the design process. Additionally,

AI can assist in selecting suitable building materials based

on environmental impact, durability, and cost, leading to

more informed decisions and greener building practices.

A traditional Building Management System (BMS)

monitors and manages building installations, such as heat-

ing, ventilation, air conditioning (HVAC), lighting, and

security systems. However, in its standard form, a BMS

operates based on predened rules and manual program-

ming, which does not qualify it as AI. A BMS can be

classied as AI when it incorporates algorithms that en-

able the system to learn from collected data and predict

optimal settings automatically. By analysing a building’s

usage patterns, external conditions, and occupancy data,

AI-powered BMS systems can dynamically adjust system

settings to meet better-changing needs. For example, AI

can automatically lower the temperature in rarely occupied

rooms, optimize energy usage based on real-time demand,

or increase ventilation during peak usage hours. These ca-

pabilities allow the BMS to operate more eciently, re-

duce energy consumption, and enhance user comfort with-

out manual intervention, distinguishing it as an intelligent

system rather than a conventional one (Ruiz et al. 2022).

The integration of machine learning in AI involves sev-

eral key steps. Initially, it is crucial to dene the AI model’s

objectives and scope and gather the necessary data for its

creation. This is followed by preliminary data processing to

address missing values, outliers, and inconsistencies. Data

pre-processing ensures consistency across dierent sourc-

es and scales, preparing the data for input into AI models.

Feature engineering involves identifying and extracting

relevant features from the data, which can be used as inputs

for the AI models. Subsequently, appropriate AI models are

selected based on the analysis goals. The model training

phase divides the dataset into training, validation, and test

sets. The AI model is trained using the training set, with ad-

justments to model parameters and architecture as needed.

The validation set is used to evaluate the model’s perfor-

mance during training, guide hyperparameter tuning, and

help prevent overtting by providing an independent data-

set for model assessment. AI is gaining increasing impor-

tance in architecture, transforming how designers approach

the creation and personalizing of buildings. Software tools

such as Midjourney, DALL-E, and Stable Diusion enable

image generation from text, becoming essential tools in the

early stages of architectural projects. These advancements

underscore the potential of AI to enhance creativity, e-

ciency, and sustainability in architectural design.

The paper delves into the selected aspects of AI’s ap-

plication in architectural design in the 21

century. It pro-

vides an understanding of how AI is personalizing archi-

tectural design, oering a glimpse into a future where AI

and human creativity work in tandem to create spaces that

are beautiful, functional, intelligent and adaptive.

State of research

AI is used in the AEC industry for planning, managing,

controlling, and personalized work, but it often neglects hu-

man-related input. The benets of human-centred AI include

personalized architectural processing, enhanced design and

engineering capabilities, data-driven project management,

improved collaboration, and increased safety.

However,

implementing human-centred AI presents challenges,

92 Małgorzata Kurcjusz, Anna Stefańska

particularly in personalizing AI systems and providing

specic training. Natural Language Processing (NLP) and

Machine Reading Comprehension (MRC) are crucial in

understanding human interests, preferences, languages,

and behaviours, facilitating better human-AI interaction

(Hu, Castro-Lacouture 2018; Nabavi et al. 2023). In archi-

tecture, AI can assist in generating design options, analys-

ing historical data for inspiration, and optimising building

performance, though current tools remain predominantly

machine-cantered. In engineering, AI supports complex

structural analyses, geotechnical engineering, smart in-

frastructure planning, and predictive maintenance, yet it

still lacks integration of human expertise. In construction,

AI aids in project planning, risk management, quality con-

trol, and predictive maintenance, but the domain faces

signicant challenges in adopting AI (Rafsanjani, Nabiza-

deh 2023).

The proposed method by Chen et al. (2023) leverages

an improved diusion model to generate high-quality ar-

chitectural designs based on textual prompts, signicant-

ly enhancing creativity and eciency in the design pro-

cess. Key highlights include the creation of a new dataset

featuring designs from eight renowned architects, which

helps train the AI model to produce designs in specic

styles. The method overcomes the limitations of tradi-

tional diusion models that struggle with specied styles

and high-quality designs. The research demonstrated that

the improved diusion model outperformed mainstream

models like Stable Diusion, Midjourney, and DALL E2,

particularly in generating designs with consistent architec-

tural styles (Kawar et al. 2023). Additionally, the study in-

troduces a set of guiding words to control design aspects,

improving the generation process’s controllability. This

AI approach promises to optimise the architectural design

workow, reducing the reliance on intensive manual la-

bour and accelerating the industry’s shift towards intelli-

gent transformation (Chen et al. 2023).

Integrating AI in education promises new perspectives,

visualisations, and interactive tools, though it also poses

challenges, such as ensuring the authenticity and reliabili-

ty of AI-generated content. The potential of AI makes the

educational process more engaging by bridging theoreti-

cal understanding and practical application through inno-

vative visualisations of historical architectural styles and

structures (Fareed, Nassif and Nofal 2024).

The three-step generation stack proposed in the re-

search by Stanislas Chaillou (2020), encompassing foot-

print massing, program repartition, and furniture layout,

exemplies a novel and sophisticated framework for ar-

chitectural design. Incorporating geographic information

system (GIS) data from the city of Boston for training

models further enhances the contextual relevance of the

generated designs. The paper provides a detailed account

of the methodology, training processes, and results, show-

casing the potential of GANs in reshaping architectural

workows through a collaborative machine-human inter-

action paradigm.

The exploration of bottom-up methods underscores their

conceptual alignment with traditional design approaches,

such as mind mapping and physical model making. T

hese

methods demonstrate adaptability through static or adap-

tive aggregations of predened building blocks, incorpo-

rating heuristic metrics like spatial relationships, environ-

mental performance, and structural eciency (Yang, Xia

2023). Additional layers during the aggregation process

oer exibility and user interaction, presenting a robust

framework for architectural exploration. Top-down meth-

ods, eective in scenarios with xed boundary constraints,

are discussed, particularly in urban design. Acknowledg-

ing their limitations in exploratory design stages, the re-

view emphasises the need for diverse representation in

datasets. It addresses concerns regarding the lack of appro-

priate curation for architectural, spatial, or cultural qual-

ities in current large-scale datasets (Weber, Mueller and

Reinhart 2022).

A proposed hierarchical framework by (Liao et al. 2024)

categorises the levels of AI integration from L0 to L5, with

current progress mainly achieving levels L1 and L2 where

AI supports human designers. Reaching higher levels,

where AI could lead or independently manage the design

process, demands additional technological advancements.

Major obstacles include the need for more high-quality

training data, the complexity of identifying and learning

from sparse features in design drawings, and the necessity

to incorporate multiple constraints into AI models. Over-

coming these

hallenges is vital for further enhancing AI

capabilities in structural design. Future research should

develop advanced data representation techniques, rene

learning algorithms to manage complex constraints, and

increase the precision and reliability of AI-generated de-

signs. Promoting open-source data sharing within the ar-

chitectural community is essential for accelerating prog-

ress and fostering innovation. This thoroughly assesses the

current state, potential advantages, and future directions

for using generative AI in building structural design (Liao

et al. 2024).

Technically, oor plan analysis research aims to con-

struct the structural model by automatically extracting

meaningful information from diverse sources. This in-

volves tasks like recognising walls and non-structural el-

ements, detecting and classifying rooms, and 2D/3D re-

construction. This interdisciplinary process spans various

disciplines within computer science, including image pro-

cessing, pattern and symbol recognition, object vectorisa-

tion, and graph modelling, highlighting the intricate nature

of oor plan analysis (Pizarro et al. 2022).

AI in architectural design optimisation

Designing assistance using AI in architecture leverag-

es advanced technologies to enhance various stages of the

design process, from conceptualisation to nal execution.

Architectural design uses AI assistance in conceptual de-

sign, planning, optimisation, and simulations (Chen 2019).

However, it is more important because it enables real-time

collaboration and automated documentation integration

in BIM environments. The AI assistance also enhances

creativity and eciency in the concept stage of architec-

tural design, improving error reduction, sustainability and

cost-eectiveness.

 TransformativeintegrationofArticialIntelligenceinarchitecturaldesign 93

Architecturalanalysis

AI in architecture enhances the design process through

advanced simulations and analyses, signicantly improv-

ing energy efficiency, structural integrity, and overall

sustainability. By simulating various environmental con-

ditions, AI helps optimise energy consumption, daylight-

ing, and thermal comfort, ensuring buildings meet green

standards and real-time interior adjustments to given basic

parameters. Structural analyses assess load-bearing capac-

ities and material performance under dierent conditions,

while life cycle and carbon footprint assessments evalu-

ate the environmental impact of building materials (Liao

et al. 2024) AI-driven acoustic and lighting simulations

enhance indoor comfort, and safety simulations optimise

evacuation routes and re safety protocols. Additionally,

urban microclimate and environmental suitability analy-

ses ensure that buildings positively interact with their

surroundings. These capabilities result in precise designs,

cost savings, and improved safety, enabling architects to

make informed decisions throughout the design process.

Foster + Partners has achieved signicant advancements

in architectural design by developing a real-time oor plan

analysis system that evaluates spatial and visual connec-

tivity (Fig. 2). These metrics are crucial for assessing the

performance of a oor plan, particularly in oce settings,

where they help determine the eectiveness of visual navi-

gation, walkability, the balance between private oces and

shared spaces, and the promotion of serendipitous collab-

oration. However, traditional methods of conducting such

analyses are time-consuming, especially for large-scale

oor plans (Tsigkari, Tarabishy, and Kosicki 2021).

The Applied Research and Development (ARD) team

at Foster + Partners aimed to reduce the workow duration

to a near real-time experience, making the analysis more

accessible and intuitive for designers during the design

process. Initially, state-of-the-art algorithms and paralleli-

sation techniques were used to reduce the analysis time to

minutes. However, this was insucient for real-time per-

formance. Consequently, the team explored machine learn-

ing to train a surrogate model capable of mimicking the

analysis output (Tsigkari, Tarabishy, and Kosicki 2021).

To achieve this, the team developed a parametric mod-

el that could generate basic oce oor plans incorporat-

ing both open-plan and compartmentalised workspaces,

complete with walls, doors, and furniture. This generative

parametric model produced a synthetic dataset of thou-

sands of oor plans. Using Hydra, the team ran spatial and

visual analyses on these synthetic plans. This resulted in

a comprehensive dataset of oor plans (inputs) and cor-

responding spatial and visual connectivity analysis results

(outputs). This dataset was then used to teach a machine

learning system, enabling it to provide spatial and visual

connectivity analyses of any oor plan without addition-

al simulation. The machine learning model continuously

rened its predictions during training by comparing its

outputs with the provided results. Initially, the model’s

predictions were inaccurate, but its accuracy improved as

it learned from the feedback. Ultimately, the model devel-

oped its method for mapping input oor plans to outputs

with correct spatial and visual connectivity analyses (Tsig-

kari, Tarabishy and Kosicki 2021).

The ecacy of this machine-learning model is illustrat-

ed through a series of images (Fig. 2). The top row displays

sample oor-plan outputs from the parametric model, in-

cluding compartmentalised and open-plan workspaces.

The middle row visualises the spatial connectivity analysis

results, while the bottom row shows the visual connectiv-

ity analysis results. Over time, the model’s predictions be-

came increasingly accurate, the model’s analysis steadily

improved and converged with the correct outputs.

Visualisations

Initially, researchers tried matching text with images

using straightforward methods. A big leap came with the

invention of Generative Adversarial Networks (GANs) in

2014, improving image quality through a generator and

a discriminator. Further advancements include the Deep

Dream program and diusion modelling, which involves

rening images from noise. Dall-E 2, created by OpenAI,

generates images from text descriptions by combining

an understanding of language and visuals. It is versatile,

user-friendly, and useful for creative tasks. Midjourney

generates images based on text inputs, which is benecial

for quickly visualising ideas, especially in art and design

(Fig. 3). Diusion Bee produces realistic images from

text descriptions, ideal for advertising and creating digital

content. It stands out for its simplicity and eectiveness.

Motionleap is a mobile app that turns text into high-quali-

ty images eortlessly and is known for its fast processing

Fig. 2. The model’s floor plan analysis based on

Tsigkari, Tarabishy, and Kosicki (2021) (source: © Foster + Partners)

Il. 2. Analiza planu piętra modelu na podstawie:

Tsigkari, Tarabishy i Kosicki (2021) (źródło: © Foster + Partners)

94 Małgorzata Kurcjusz, Anna Stefańska

and ease of use. Dall-E, Midjourney, Diusion Bee, and

Motionleap each have distinct methods for processing text

descriptions and dierent user interfaces, oering a vari-

ety of visual creation tools. The suitability of these tech-

nologies for architectural visualisation varies, with each

having specic features and technical abilities that aect

their utility. Examining Midjourney queries highlights

common architectural visualisation trends, demonstrating

how AI tools facilitate creative exploration and design

(Hanafy 2023). Text-to-image generation uses machine

learning methods, specically generative models, to trans-

form textual descriptions into visually appealing images

that closely match the semantic intent of the text. These

models, which include Autoencoders, GANs, Deep Boltz-

mann Machines, and Diusion models, learn to create new

images by analysing vast training data. Diusion models

stand out for their eciency in image creation by gradual-

ly adding and then removing noise to form the nal image.

This process involves several innovative techniques to un-

derstand and process textual inputs, including tokenisation

and embedding (Fareed, Nassif and Nofal 2024). Gener-

ative models nd applications across various elds, from

creating images based on text descriptions to enhancing

computer vision and image recognition capabilities. The

conditioning technique, crucial in these models, guides the

noise reduction based on textual descriptions facilitated

by advanced tokenisation and embedding methods. These

steps convert text into numerical values that the model can

understand, allowing for generating contextually relevant

images (Fareed, Nassif and Nofal 2024).

Recent advancements in AI art generation tools are sig-

nicantly enhancing the capabilities of architects and de-

signers, capturing the attention of professionals worldwide.

The paper by Ploennigs and Berger (2023) delves into AI’s

current and potential applications in architecture, comparing

the technologies behind three leading AI art platforms, an-

alysing their use cases, and sharing practical workows de-

rived from extensive experience with these systems. Gener-

ative AI models, especially text-to-image generators, are at

the core of these platforms, leveraging GPT-3 for generat-

ing text and specialised versions for predicting image pixels

from textual inputs. Diusion models, evolving from ini-

tial proposals to more sophisticated systems like OpenAI’s

GLIDE and DALL-E, play a crucial role by generating new

images through a reverse diusion process from noise-add-

ed images. DALL-E, Stable Diusion, and Midjourney

support diverse image generation tasks from user prompts,

including image modication,

inpainting, outpainting, and

super-resolution (Ploennigs, Berger 2023). It is essential to

employ encoding methods that eectively capture the intri-

cate details of topology, geometry, spatial information, and

other relevant aspects of building structure design. When

these encoding methods are paired with suitable generative

AI algorithms, they can signicantly enhance the generation

of structural designs. However, it is crucial to acknowledge

that no single method can capture all design features fully,

necessitating a focus on multi-modal data fusion to improve

understanding and representation of building designs,

a general representation technique to address the sparsity

of architectural information, and the importance of sharing

open-source data within the architectural community to fos-

ter innovation (Liao et al. 2024).

The advent of AI in architectural visualisation has revo-

lutionised the speed and eciency with which visual rep-

resentations of designs can be created. AI-driven visualisa-

tion tools, such as Adobe Firey, enable rapid generation

and modication of graphics based on input data, wheth-

er textual descriptions or existing images (Fig. 3). This

case study explores the capabilities and limitations of AI

in the context of architectural visualisations. AI-powered

Fig. 3. Images generated in AI software Adobe Firefly (elaborated by M. Kurcjusz)

Il. 3. Obrazy wygenerowane w programie AI Adobe Firefly (oprac. M. Kurcjusz)

 TransformativeintegrationofArticialIntelligenceinarchitecturaldesign 95

visualisations can produce detailed and accurate images

quickly, signicantly reducing the time required to create

high-quality visual content. These tools allow for extensive

customisation, enabling users to adjust lighting, viewing

angles, styles, eects, and reference materials to achieve

the desired aesthetic. The process begins with inputting

data, which can be descriptive text or pre-existing images.

The AI then processes this input to generate visualisations

that align with the specied criteria. This capability allows

designers to quickly explore multiple design options and

variations, facilitating a more iterative and exible de-

sign process. However, despite its advanced capabilities,

AI-driven visualisation technology has limitations. One of

the primary challenges is that the AI may not always pro-

duce results that perfectly match the user’s expectations.

While AI can generate highly detailed and contextually

appropriate images, achieving the desired outcome may

require additional manual adjustments and iterations. This

limitation underscores the importance of human expertise

in guiding and rening the outputs generated by AI. More-

over, as AI technology evolves, enhancing its ability to in-

terpret and execute complex design instructions accurately

remains an ongoing challenge. AI tools are procient at

handling straightforward tasks and generating initial design

concepts. However, they still require signicant input and

oversight from human designers to ensure that the nal out-

puts meet the nuanced demands of architectural projects.

SpectrumofAIapplicationsinarchitecture

By leveraging advanced tools and algorithms, architects

and engineers can address complex challenges, streamline

processes, and deliver innovative solutions that meet mod-

ern needs. Table 1 provides an overview of the spectrum

Category AI application Work performed Effectiveness assessment Examples of tools

Concept design

Generative Design

generates multiple

optimized design options

based on constraints

like materials and cost

reduces design time,

provides innovative

and optimized solutions

Autodesk Fusion

(generative design AI),

Hypar, Autodesk Forma

Parametric Design

creates dynamic, rule-based

design variations

for complex forms

facilitates precision

and flexibility in complex

geometric design

Rhino Grasshopper

(AI Plugins)

Building

performance

Energy efficiency

optimization

virtual reconstruction

of heritage sites

reduces energy costs

and improves building

performance

Autodesk Forma

Simulations

(thermal, daylight,

noise, wind)

simulates thermal comfort,

daylighting, and energy

efficiency

enhances comfort

and sustainability

while ensuring regulatory

compliance

ClimateStudio

(AI-assisted analysis),

Autodesk Forma

Structural integrity

analysis

assesses structural stability

and material performance

minimizes structural costs

and improves safety by

predicting failure risks

T2D2 (AI Structural

Damage Detection)

Design automation

Layout optimization

automates the creation

of optimized floor plans

and layouts

reduces design time

and improves spatial

efficiency

ArchiGAN, Autodesk

Fusion, Hypar

Code compliance checking

ensures compliance with

local building regulations

reduces human error,

ensuring regulatory

alignment and avoiding

costly revisions

Bamroc

Material

and resource use

Smart Material selection

selects sustainable,

cost-effective building

materials

reduces environmental

impact and optimizes costs

Cove.Tool

(AI for sustainable design)

Resource and waste

optimisation

minimizes material usage

and waste during

design/construction

cuts material costs and

reduces waste significantly

SmartWaste

Construction

management

Scheduling and resource

allocation

predicts timelines,

allocates labour/resources

effectively

improves project delivery

by reducing delays and

overruns

ALICE Technologies,

nPlan

Site safety monitoring

detects safety risks

in real-time using AI

and computer vision

reduces accidents on-site,

enhancing worker safety

Smartvid.io

Table 1. Spectrum of AI applications in architecture (elaborated by M. Kurcjusz)

Tabela 1. Spektrum zastosowań AI w architekturze (oprac. M. Kurcjusz)

96 Małgorzata Kurcjusz, Anna Stefańska

of AI applications in architecture, including examples of

tools, the specic work performed, and an assessment of

their eectiveness.

Discussion

Integrating AI in architectural design has demonstrat-

ed signicant potential to revolutionise the eld. This

study has explored various dimensions of AI applications,

including generative design, optimisation, and predic-

tive analytics. The ndings suggest that AI can enhance

architectural practices’ creativity, eciency, and sustain-

ability. AI-driven generative design tools have shown the

ability to produce innovative and diverse design solutions

that might not be conceived through traditional methods.

These tools leverage algorithms to explore a vast design

space, oering architects many options. This fosters cre-

ativity and allows for exploring unconventional design

forms and structures. The case studies reviewed in this

paper highlight successful implementations where AI has

contributed to groundbreaking architectural projects. AI

optimisation applications have proven eective in stream-

lining various aspects of the design process.

The ndings from this study highlight a typology of

AI applications in architectural design, categorizing their

contributions into conceptual design, performance optimi-

zation, and sustainability enhancement. At the conceptu-

al level, AI-driven generative and parametric design tools

have proven instrumental in producing diverse and inno-

vative design solutions. These tools allow architects to ex-

plore vast design spaces, fostering creativity while reducing

manual eort. At the performance optimization level, AI

applications, such as structural analysis and energy simula-

tions, demonstrated signicant improvements in eciency

and precision. The experimental results revealed a marked

reduction in design iterations, material waste, and ener-

gy consumption, underscoring AI’s potential to optimize

critical aspects of the design and construction process. At

Category AI application Work performed Effectiveness assessment Examples of tools

BIM integration

Data analysis

and prediction

AI analyses BIM data

to predict workflows

and maintenance

optimizes project timelines

and reduces lifecycle costs

Spacio AI

Generative BIM Models

automates creation

of intelligent BIM models

improves efficiency

and accuracy during

design phases

Revit Dynamo

(AI-Integrated),

BricsCAD BIM, WiseBIM

Sustainability

and climate

Carbon footprint

assessment

evaluates environmental

impact of materials

and design choices

enables reductions in

embodied carbon

One Click LCA

(AI-based analysis),

Autodesk Forma

Climate-responsive design

adapts building design

to local climate

for efficiency

reduces operational energy

consumption and improves

occupant comfort

Autodesk Forma

Post-Occupancy

management

Smart Building

Management Systems

(BMS)

learns from real-time data

to adjust HVAC, lighting,

and security

reduces energy use while

improving building

comfort

Siemens Desigo CC

Predictive Maintenance

predicts equipment failures

and schedules proactive

maintenance

cuts maintenance costs

and downtime

IBM Maximo Manage AI,

Schneider EcoStruxure

Visualissation

and design

AI-Powered Rendering

Tools

creates high-quality,

realistic visualizations

of designs

reduces visualization time

while improving quality

and stakeholder

communication

MidJourney, DALL-E 2,

Stable Diffusion

Text-to-Image Generation

generates conceptual

imagery from text prompts

enhances creativity

and reduces early design

iteration times

Adobe Firefly,

MidJourney

Human-cantered

design

User Experience Analysis

analyses behaviour

to optimize user-centric

designs

improves functional

efficiency and user

satisfaction in built

environments

Autodesk Forma

Space Utilization

Optimization

optimizes interior layouts

based on occupancy data

increases spatial efficiency

and reduces operational costs

Density.io AI

Historical

preservation

Virtual Reconstruction

of Heritage Sites

reconstructs historical sites

using AI-driven

3D modelling

facilitates accurate digital

preservation

and restoration planning

RealityCapture AI

Table 1 continued. Spectrum of AI applications in architecture (elaborated by M. Kurcjusz)

Tabela 1 cd. Spektrum zastosowań AI w architekturze (oprac. M. Kurcjusz)

 TransformativeintegrationofArticialIntelligenceinarchitecturaldesign 97

the sustainability level, integrating predictive analytics and

simulation models facilitated the selection of environmen-

tally friendly materials and the achievement of green build-

ing standards, eectively reducing carbon footprints. AI

algorithms can process large datasets and perform com

plex

calculations at a speed and accuracy unattainable by human

designers alone. This leads to more ecient use of

materi-

als, cost savings, and reduced project timelines. Integrat-

ing AI in BIM further exemplies how AI can enhance

coordination and collaboration among different stake-

holders in a project. Presented levels of AI applications

collectively illustrate how AI transitions from enhancing

creativity to ensuring technical precision and promot-

ing environmental stewardship in architec tural design.

Adopting AI in architectural design introduces signif-

icant benets, such as enhanced automation, improved

precision, and rapid optimization. However, AI’s hasty

and unregulated application also poses substantial risks

that must be carefully considered. AI algorithms often rely

on historical data and existing patterns, which can lead to

the replication of safe solutions and hinder the creation of

unique, innovative designs. Over-reliance on AI tools risks

diminishing the creative role of architects, reducing their

ability to think outside the box and limiting human inge-

nuity in favour of algorithm-driven recommendations. Ad-

ditionally, many AI models, particularly neural networks,

operate as black boxes, making it dicult to trace how

a design decision was reached. This lack of explainability

raises concerns about error detection and accountability,

particularly when errors go unnoticed until later stages, po-

tentially leading to costly rework or safety compromises.

Despite these advancements, the reliance on high-qual-

ity data, the need for specialized expertise, and ethical

considerations regarding decision-making accountability

remain key challenges to address. The hasty application

of AI in architectural design brings potential risks. The

quality of AI outputs depends heavily on the data used to

train the models. Biased, incomplete, or low-quality data-

sets can result in awed or unsafe designs that fail to meet

user requirements or safety standards. AI tools may also

struggle to comply with local building codes, zoning laws,

or regulations, increasing the risk of non-compliant de-

signs and potential legal and nancial repercussions (Al-

kfairy et al. 2024). Moreover, questions about account-

ability arise during design failures: determining whether

responsibility lies with the architect, software provider, or

AI developer remains a legal and ethical challenge. Anoth-

er notable risk is AI’s inability to capture subtle cultural,

historical, or social nuances, often integral to architectural

projects. Overlooking these factors can result in designs

that lack cultural resonance or fail to align with user ex-

pectations. Furthermore, the high costs of implementing

AI and the need for specialized skills can act as barriers to

adoption, particularly for smaller rms. Ethical concerns,

such as job displacement and the potential erosion of hu-

man judgment in decision-making, further complicate AI’s

integration into architectural workows. While AI holds

transformative potential for architectural design, its limita-

tions and associated risks must not be ignored. Achieving

a balanced, human-AI collaborative approach with clear

regulatory frameworks, accountability measures, and ro-

bust ethical standards will ensure that AI enhances archi-

tectural innovation while minimizing its challenges.

Sustainability is a critical concern in contemporary ar-

chitecture, and AI has shown promise in addressing this

challenge. Predictive analytics powered by AI can assess

the environmental impact of design choices, optimise

energy consumption, and suggest sustainable materials.

The study’s ndings indicate that AI can play a pivotal

role in achieving green building standards and reducing

the carbon footprint of construction projects. Despite the

promising potential, integrating AI in architectural design

is challenging. One signicant concern is the reliance on

high-quality data for AI algorithms to function eective-

ly. Inaccurate or incomplete data can lead to suboptimal

design outcomes. Additionally, architects need to acquire

new skills to work eectively with AI tools, which may

require substantial training and adaptation.

Conclusion

The transformative integration of AI in architectural

design holds immense potential to reshape the industry.

AI can contribute to more innovative, cost-eective, and

environmentally friendly architectural solutions by en-

hancing creativity, improving eciency, and promoting

sustainability. However, addressing the challenges related

to data quality, skill acquisition, and ethical implications

is essential to realise these benets fully. Future research

should focus on developing robust AI frameworks that

seamlessly integrate with existing architectural practices

and creating educational programs to equip architects with

the necessary skills to leverage AI technologies eectively.

While AI enables multi-criteria analysis to evaluate

and optimize various project aspects – such as cost, ener-

gy eciency, aesthetics, and environmental impact – its

introduction into such a complex and nuanced process is

not without diculties. One of the primary challenges lies

in the complexity and ambiguity of architectural criteria.

Architecture encompasses technical, aesthetic, social,

and environmental dimensions, many of which are di-

cult to quantify or dene as clear inputs for AI systems.

Non-measurable values, such as subjective aesthetics or

the project’s impact on the local community, remain chal-

lenging for AI to interpret, as concepts like “beauty” or

“harmony with the environment” are inherently subjective

and inuenced by cultural context.

Another major challenge is the quality of input data. Ar-

chitectural data is often incomplete, inaccurate, or unstruc-

tured, which poses signicant obstacles for AI algorithms

that rely on high-quality data to deliver reliable results.

Furthermore, AI must address conicting criteria within

the design process – for instance, achieving low construc-

tion costs may conict with goals of high energy eciency,

aesthetic appeal, or sustainability. Without clearly dened

priorities, AI may favour one criterion, such as cost mini-

mization, while neglecting others that are equally import-

ant, such as environmental considerations (Kurcjusz 2024).

Additional challenges stem from the limited adaptation

of AI algorithms to the unique requirements of architecture.

98 Małgorzata Kurcjusz, Anna Stefańska

References

Abdulfattah, Basem S., Hassan A. Abdelsalam, Mai Abdelsalam, Mar-

zia Bolpagni, Niraj Thurairajah, Laura Florez Perez, and Talib E.

Butt. “Predicting Implications of Design Changes in BIM-Based

Construction Projects through Machine Learning.” Automationin

Construction 155 (2023): 105057. https://doi.org/10.1016/j.aut-

con.2023.105057.

Abioye, Soat O., Lukumon O. Oyedele, Lukman Akanbi, et al. “Arti-

 cial Intelligence in the Construction Industry: A Review of Pre-

sent Status, Opportunities and Future Challenges.” Journal of

BuildingEngineering 44 (2021): 103299. https://doi.org/10.1016/j.

jobe.2021.103299.

Al-kfairy, Mousa, Dheya Mustafa, Nir Kshetri, Mazen Insiew, and Omar

Alfandi. “Ethical Challenges and Solutions of Generative AI: An

Interdisciplinary Perspective.” Informatics 11, no. 3 (2024): 58.

https://doi.org/10.3390/informatics11030058.

Chaillou, Stanislas. “ArchiGAN: Articial Intelligence x Architecture.”’

In ArchitecturalIntelligence, edited by Philip F. Yuan, Mike Xie,

Neil Leach, Jiawei Yao, Xiang Wang. Springer Nature Singapore,

2020. https://doi.org/10.1007/978-981-15-6568-7_8.

Chen, Junming, Duolin Wang, Zichun Shao, Xu Zhang, Mengchao

Ruan, Huiting Li, and Jiaqi Li. “Using Articial Intelligence to

Generate Master-Quality Architectural Designs from Text Descrip-

tions.” Buildings 13, no. 9 (2023): 2285. https://doi.org/10.3390/

buildings13092285.

Chen, Shang Yuan. “Enhancing Validity of Green Building Information

Modeling with Artificial-Neural-Network-Supervised Learning

– Taking Construction of Adaptive Building Envelope Based on

Daylight Simulation as an Example.” SensorsandMaterials 31,

no. 6 (2019): 1831–45. https://doi.org/10.18494/SAM.2019.2147.

Elwy, Ibrahim, and Aya Hagishima. “The Articial Intelligence Ref-

ormation of Sustainable Building Design Approach: A Systematic

Review on Building Design Optimization Methods Using Surrogate

Models.” EnergyandBuildings 323, (2024): 114769. https://doi.

org/10.1016/j.enbuild.2024.114769.

Fareed, Mohamed W., Ali Bou Nassif, and Eslam Nofal. “Exploring the

Potentials of Articial Intelligence Image Generators for Educating

the History of Architecture.” Heritage 7, no. 3 (2024): 1727–53.

https://doi.org/10.3390/heritage7030081.

Ganiyu, Sikiru Abiodun, Lukumon O. Oyedele, Olugbenga Akinade,

Hakeem Owolabi, Lukman Akanbi, and Abdulqayum Gbadamo-

si. “BIM Competencies for Delivering Waste-Ecient Building

Projects in a Circular Economy.” Developments in the BuiltEn-

vironment 4 (2020): 100036. https://doi.org/10.1016/j.dibe.2020.

100036.

Hanafy, Nervana Osama. “Articial Intelligence’s Eects on Design

Process Creativity: ‘A Study on Used A.I. Text-to-Image in Ar-

chitecture’.” JournalofBuildingEngineering 80 (2023): 107999.

https://doi.org/10.1016/j.jobe.2023.107999.

Hu, Yuqing, and Daniel Castro-Lacouture. “Clash Relevance Prediction

Based on Machine Learning.” JournalofComputinginCivilEn-

gineering 33, no. 2 (2018): 1–15. https://doi.org/10.1061/(ASCE)

CP.1943-5487.0000810.

Kawar, Bahjat, Shiran Zada, Oran Lang, et al. “Imagic: Text-Based

Real Image Editing with Diusion Models.” 2023Proceedingsof

theIEEEComputerSocietyConferenceonComputerVisionand

Pattern Recognition, Vancouver, Canada (2023): 6007–17. doi:

10.1109/CVPR52729.2023.00582.

Khan, Ayaz Ahmad, Abdulkabir Opeyemi Bello, Mohammad Arqam,

and Fahim Ullah. “Integrating Building Information Modelling and

Articial Intelligence in Construction Projects: A Review of Chal-

lenges and Mitigation Strategies.” Technologies 12, no. 10 (2024):

185. https://doi.org/10.3390/technologies12100185.

Kineber, Ahmed Farouk, Nehal Elshaboury, Ayodeji Emmanuel Oke, John

Aliu, Ziyad Abunada, and Mohammad Alhusban. “Revolutionizing

Construction: A Cutting-Edge Decision-Making Model for Arti-

cial Intelligence Implementation in Sustainable Building Projects.”

Heliyon 10, no. 17 (2024): e37078. https://doi.org/10.1016/j.heliy-

on.2024.e37078.

Kurcjusz, Małgorzata. “Technology vs Creativity in Architecture: Striv-

ing for Synergy.” In DeningtheArchitecturalSpaceArchitecture

andTechnology. Vol.4, edited by Anna Mielnik, Ocyna Wydawni-

cza ATUT – Wrocławskie Wydawnictwo Oświatowe, 2024. https://

doi.org/10.23817/2024.defarch.4-2.

Liao, Wenjie, Xinzheng Lu, Yifan Fei, Yi Gu, and Yuli Huang. “Gener-

ative AI Design for Building Structures.” AutomationinConstruc-

tion 157 (2024): 105187. https://doi.org/10.1016/j.autcon.2023.

105187.

Nabavi, Armin, Issa Ramaji, Naimeh Sadeghi, and Anne Anderson. “Le-

veraging Natural Language Processing for Automated Information

Inquiry from Building Information Models.” JournalofInforma-

tionTechnologyin Construction 28 (2023): 266–85. https://doi.

org/10.36680/J.ITCON.2023.013.

Pizarro, Pablo N., Nancy Hitschfeld, Ivan Sipiran, and Jose M. Saave-

dra. “Automatic Floor Plan Analysis and Recognition.” Automation

inConstruction 140 (2022): 104348. https://doi.org/10.1016/j.aut-

con.2022.104348.

Ploennigs, Joern, and Markus Berger. “AI Art in Architecture.” AIin

CivilEngineering 2, (2023): 8. https://doi.org/10.1007/s43503-023-

00018-y.

Rafsanjani, Hamed Nabizadeh, and Amir Hossein Nabizadeh. “Towards

Human-Centered Articial Intelligence (AI) in Architecture, Engi-

neering, and Construction (AEC) Industry.” ComputersinHuman

Behavior Reports 11 (2023): 100319. https://doi.org/10.1016/j.

chbr.2023.100319.

Ruiz, M. Dolores, Juan Gomez-Romero, Carlos Fernandez-Basso, and

Maria J. Martin-Bautista. “Big Data Architecture for Building

Energy Management Systems.” IEEE Transactionson Industri-

alInformatics 18, no. 9 (2022): 5738–47. https://doi.org/10.1109/

TII.2021.3130052.

Tsigkari, Martha, Sherif Tarabishy, and Marcin Kosicki. “Towards Arti-

cial Intelligence in Architecture: How Machine Learning Can Change

the Way We Approach Design.” PlusJournal. Published March 29,

2021. Accessed February 5, 2025, at https://www.fosterandpartners.

com/insights/plus-journal/towards-articial-intelligence-in-architec-

ture-how-machine-learning-can-change-the-way-we-approach-de-

sign.

Weber, Ramon Elias, Caitlin Mueller, and Christoph Reinhart. “Au-

tomated Floorplan Generation in Architectural Design: A Review

of Methods and Applications.” Automationin Construction 140

(2022): 104385. https://doi.org/10.1016/j.autcon.2022.104385.

Yang, Honglei, and Mim Xia. “Advancing Bridge Construction Mon-

itoring: AI-Based Building Information Modeling for Intelligent

Structural Damage Recognition.” AppliedArticialIntelligence 37,

no. 1 (2023). https://doi.org/10.1080/08839514.2023.2224995.

Elements such as a building’s impact on the surrounding

landscape, spatial relationships, or compliance with historic

preservation regulations are dicult for standard AI models

to capture. For example, an algorithm optimized to reduce

material costs may fail to consider aesthetic requirements or

cultural heritage constraints.

In conclusion, while AI presents exciting opportunities

for the future of architectural design, a balanced approach

that considers both the advantages and the challenges is

crucial for its successful implementation. The ongoing

collaboration between AI experts and architects will be

key to unlocking the full potential of AI in this eld.

 TransformativeintegrationofArticialIntelligenceinarchitecturaldesign 99

Streszczenie

Transformacyjna integracja sztucznej inteligencji w projektowaniu architektonicznym

Wprowadzenie sztucznej inteligencji (Articial Intelligence – AI) do projektowania architektonicznego zmienia tradycyjne praktyki. W artykule

przedstawiono typologiczne zastosowania AI w architekturze, koncentrując się na jej wkładzie w projektowanie koncepcyjne, optymalizację wy-

dajności oraz zrównoważony rozwój, podkreślając jej transformacyjny wpływ na procesy w sektorze architektury, inżynierii i budownictwa (AEC).

Celem autorek pracy było sklasykowanie i analiza hierarchicznych zastosowań AI w projektowaniu architektonicznym. Głównym zamie-

rzeniem – ocena roli AI w rozwijaniu kreatywności, zwiększaniu efektywności i wspieraniu praktyk zrównoważonego rozwoju. Metodologicznie

badanie oparto na kompleksowym przeglądzie klasykacji typologicznych, walidacji eksperymentalnych oraz analizach studiów przypadków. Syn-

teza wyników z najnowszej literatury przedmiotu i rzeczywistych implementacji pozwala na ocenę potencjału i ograniczeń AI na różnych etapach

projektowania architektonicznego.

Wyniki badania wskazują na hierarchię typologiczną zastosowań AI w projektowaniu architektonicznym, ukazując jej wielowymiarowy wkład.

W projektowaniu koncepcyjnym narzędzia generatywne napędzane AI umożliwiają architektom eksplorację rozległych przestrzeni projektowych,

pozwalając na tworzenie innowacyjnych i różnorodnych rozwiązań. W zakresie optymalizacji wydajności AI wykazuje znaczną poprawę efektyw-

ności poprzez redukcję czasu iteracji projektowych, minimalizację marnotrawstwa materiałów oraz poprawę efektywności energetycznej dzięki

zaawansowanym symulacjom. W dziedzinie zrównoważonego rozwoju analityka predykcyjna AI wspiera podejmowanie świadomych decyzji do-

tyczących wyboru materiałów oraz zgodności ze standardami ekologicznymi, przyczyniając się do uzyskania certykatów zielonego budownictwa.

Mimo tych postępów wyzwaniami pozostają zależność od wysokiej jakości danych, kwestie etyczne związane z odpowiedzialnością za decyzje oraz

potrzeba specjalistycznych kompetencji. Na podstawie otrzymanych wyników można stwierdzić, że AI wzmacnia kreatywność, precyzję i zrówno-

ważony rozwój w architekturze, ale jej skuteczna integracja wymaga współpracy człowieka i AI, wspieranej przez solidne ramy etyczne oraz rozwój

kompetencji.

Słowa kluczowe: sztuczna inteligencja, algorytmy, parametryka, współczesne projektowanie architektoniczne