Content uploaded by James Hutson
Author content
All content in this area was uploaded by James Hutson on Jan 23, 2024
Content may be subject to copyright.
Forum for Art Studies 2024, 1(1), 393.
https://doi.org/10.59400/fas.v1i1.393
1
Article
Integrating art and AI: Evaluating the educational impact of AI tools in
digital art history learning
James Hutson
Department Head of Art History and Visual Culture, College of Arts and Humanities, Lindenwood University, Saint Charles, MO 63301, USA;
jhutson@lindenwood.edu
Abstract: This study delves into the burgeoning intersection of Artificial Intelligence (AI)
and art history education, an area that has been relatively unexplored. The research focuses
on how AI art generators impact learning outcomes in art history for both undergraduate
and graduate students enrolled in Ancient Art courses, covering eras from ancient
Mesopotamia to the fall of Rome. Utilizing a mixed-methods approach, the study analyzes
AI-generated artworks, reflective essays, and survey responses to assess how these
generative tools influence students’ comprehension, engagement, and creative
interpretation of historical artworks. The study reveals that the use of AI tools in art history
not only enhances students’ understanding of artistic concepts but also fosters a deeper,
more nuanced appreciation of art from the periods studied. The findings indicate that
engaging with AI tools promotes critical thinking and creativity, which are crucial
competencies in the study of art history. Survey data further suggest that the integration of
AI in art history positively influences students’ perceptions of the discipline, aligning well
with contemporary digital trends. One of the significant outcomes of the study is the varied
experiences of students with AI tools. While some faced challenges with the technology,
particularly in accurately capturing complex artwork details and crafting effective
prompts, others found success in using AI to generate detailed and creative interpretations
of historical pieces. These experiences underscore the potential of AI as a valuable
pedagogical tool in art history and humanities education, offering novel insights into
teaching methodologies.
Keywords: digital art history; formal analysis; close looking; AI art; generative AI
1. Introduction
Art history has traditionally been taught through a combination of lectures and
discussion, focusing heavily on visual analysis and close looking [1]. The approach,
often termed formal analysis, centers on the detailed examination of the visual features
of artworks and involves describing these features systematically and analyzing their
effects, using a set of established terms and concepts like format, scale, composition,
viewpoint, treatment of the human figure and space, and the use of form, line, color,
light, and texture [2]. This method aims to decipher and convey the visual experience
of art, guiding students to understand not just the subject matter or historical context,
but the artistic form itself and its impact on viewers. Over time, while the essence of
formal analysis remains, its interpretation has shifted from a presumed universality in
human response to visual form to a more subjective understanding, still valued as a
critical exercise in analyzing visual experience, especially in introductory art history
courses [3].
CITATION
Hutson J. Integrating art and AI:
Evaluating the educational impact of
AI tools in digital art history learning.
Forum for Art Studies. 2024; 1(1):
393.
https://doi.org/10.59400/fas.v1i1.393
ARTICLE INFO
Received: 8 December 2023
Accepted: 19 December 2023
Available online: 23 January 2024
COPYRIGHT
Copyright © 2024 author(s).
Forum for Art Studies is published by
Academic Publishing Pte. Ltd. This
article is licensed under the Creative
Commons Attribution License (CC
BY 4.0).
http://creativecommons.org/licenses/
by/4.0/
Forum for Art Studies 2024, 1(1), 393.
2
In teaching formal analysis, art history educators often employ a variety of
assignments designed to hone skills in close-looking and critical thinking. A common
assignment is a close-looking exercise, where students are asked to spend an extended
period observing a single artwork, noting every detail, from the brushwork to the use
of light and color [4]. These exercises train them to notice nuances they might
otherwise overlook and develop a deeper understanding of artistic techniques. Another
typical assignment is the compare and contrast essay, where students select two
artworks, often from different periods or styles, and analyze their similarities and
differences in form, technique, and content [5]. The task encourages students to think
critically about how different contexts and artistic movements affect visual expression,
while the outcome of such analysis should be a description that effectively verbalizes
the visual experience. This means converting visual stimuli into words in such a way
that a reader can recreate a mental picture of the artwork being described. Such
descriptions call attention to important qualities of the work, such as composition, use
of color and light, thematic elements, and stylistic nuances. The ability to produce such
vivid, detailed descriptions is a fundamental skill in art history, enabling students to
communicate their visual experiences and insights effectively [6].
While traditional approaches to improving student abilities in formal analysis
have remained fairly analog, more digital interventions are being seen [7]. The
increasing integration of emergent technologies in digital art history has been
spearheaded by innovations such as virtual reality (VR), which has notably enhanced
formal analysis skills. VR provides an immersive environment that allows for detailed
comparisons and replicates the experience of in-person art viewing, thus significantly
contributing to the development of critical analytical skills in art history [8]. However,
the potential of new abilities of generative artificial intelligence (AI) as a
complementary tool in this domain has yet to be fully explored. AI, especially in the
form of AI art generators, offers a unique opportunity to further refine formal analysis
skills. By enabling students to recreate historical artworks using AI, they engage in a
deep, hands-on exploration of art elements such as composition, style, and technique.
This approach not only reinforces their understanding of art history but also introduces
them to the innovative possibilities of AI in art creation and study. Consequently, as
Zullich et al. [9] have observed, AI stands as a promising addition to the digital art
history classroom, complementing existing technologies like VR and broadening the
scope of teaching methodologies in art historical analysis.
Aligning these traditional strategies with prompt engineering for generative
visual models of AI, such as Dall-E 3, Midjourney, or Stable Diffusion, offers an
intriguing extension of this skillset. In the context of AI, prompt engineering involves
crafting descriptive prompts to guide AI in generating specific visual outputs. This
process closely mirrors formal analysis in art history; both require a keen eye for detail
and the ability to translate visual elements into precise, descriptive language.
However, in the case of these new generative visual tools, the descriptions are used as
inputs for the model to create visual representations [10]. By applying their formal
analysis skills to prompt engineering, students can learn to manipulate generative AI
tools effectively, crafting prompts that result in accurate visual recreations of historical
artworks. The intersection of traditional art historical techniques and modern AI
Forum for Art Studies 2024, 1(1), 393.
3
technology represents a novel and potentially powerful tool in art education, blending
close-looking and descriptive prowess with digital innovation [11].
The strategies of teaching formal analysis and close looking at art history, while
well-established, have not been extensively explored in the context of modern AI
technologies, particularly in the realm of prompt engineering with AI generative tools.
This study represents a pioneering effort to investigate how these traditional art
historical methodologies can be adapted and applied in an AI-centric environment.
The pedagogical strategy devised for this study involves an innovative approach:
starting with the traditional methods of teaching formal analysis and close looking,
students are then tasked to use these skills for prompt engineering. The objective is to
recreate, as closely as possible, the original artwork being described using an AI
generative tool, such as Dall-E 3 or Midjourney. A crucial aspect of this approach is
the restriction placed on students from using identifiable descriptors in their prompts,
such as the title, region, style, or any other metadata of the artwork that the training
model might scrape from. This constraint ensures that the students rely solely on their
ability to verbalize the visual elements and qualities of the artwork without leaning on
easily identifiable, pre-categorized information.
The methodology employed in this study, encompassing the analysis of AI-
generated artworks, reflective essays, and survey responses, is designed to thoroughly
evaluate the effectiveness of integrating AI tools in art history education. The
anticipated results are expected to shed light on the depth of understanding and level
of engagement students attain through this novel pedagogical approach. By closely
examining the outcomes of using AI art generators, the study aims to discern how
these tools influence the development of formal analysis skills and overall learning
experience in art history. Expected outcomes include insights into how students
perceive and interact with AI technology in an academic setting, the challenges they
encounter, and the successes they achieve. The study also aims to explore how AI-
generated images and the process of creating them enhance students’ understanding
of historical artworks. The results are anticipated to provide a nuanced understanding
of the potential role of AI tools in art history education, highlighting both the
opportunities and limitations of this approach.
Based on the data and discussions in this study, the findings indeed offer
significant contributions to the pedagogy of art history. The results indicate that while
AI tools, such as Stable Diffusion and Dalle-3, have the potential to enhance the
teaching and learning of art history, particularly in the areas of formal analysis and
close looking, their integration requires careful consideration. The study showed that
students’ experiences with AI tools varied: while some found these tools effective in
enhancing their understanding of artworks, others faced challenges with anatomical
accuracy and prompt specificity. The findings also revealed a diversity in students’
perceptions of AI’s role in art history education. While some students recognized the
potential of AI in providing novel perspectives and aiding in visual analysis, others
remained skeptical about its effectiveness and expressed concerns about AI’s
limitations. These results underscore the need for a balanced approach to AI
integration, one that acknowledges both its capabilities and its limitations. Ultimately,
this study suggests that emerging technologies like AI can be innovatively integrated
into traditional teaching methodologies in art history, potentially transforming
Forum for Art Studies 2024, 1(1), 393.
4
teaching practices. However, this integration should be done thoughtfully, ensuring
that technology complements rather than replaces traditional methods. The
implications of this study extend beyond art history, providing a model for how
technology can be integrated into other humanities disciplines, thereby contributing to
the broader educational landscape. This integration, if done effectively, can enrich
learning experiences and open up new avenues for exploration and understanding in
the humanities.
2. Literature review
Teaching formal analysis and close looking in art history classes is pivotal for
cultivating an in-depth understanding of art history and the intricacies of individual
artworks. This pedagogical approach emphasizes detailed observation and the
articulation of visual elements, which are central to formal analysis. Walton [12]
highlights the significance of finding the right words to describe visual aspects, an
essential skill in art historical analysis. According to Gasper-Hulvat [13], formal
analysis, which examines form and structure of a work, is often the foundation of art
historical inquiry and serves as a critical method for discussing art, particularly when
integrated with other interpretative approaches. This type of analysis typically
involves a thorough visual examination of art, focusing on its constituent elements and
the principles of design employed by artists across different mediums.
Within art history courses, students engage with a variety of assignments that
encourage these analytical skills. These tasks range from visual analysis essays, where
students focus intensely on the artwork itself, to comparison essays and research
papers that may incorporate primary textual sources for deeper contextual
understanding [14]. The emphasis in visual analysis essays is primarily on close
observation, often minimizing the reliance on extensive secondary sources [15]. Such
assignments are designed to hone student abilities to scrutinize and articulate their
observations, thereby deepening their understanding of art historical methods and the
artworks themselves. Successful navigation of these tasks requires a clear
comprehension of the assignment prompts, underscoring the importance of aligning
analytical skills with specific academic expectations [16].
The thematic exploration of formal analysis scholarship in art history reveals its
multifaceted role and evolving methodologies. For instance, Olin [17] delves into the
historical roots of formal analysis in the United States, tracing its origins to the
nineteenth century. Initially perceived as a detached method for classifying and
analyzing art, drawing inspiration from natural sciences, formal analysis was
considered foundational to art historical thought. However, Olin challenges this view,
suggesting that it was more of a superstructure, with its authority lending it a
foundational appearance. The study also explores the interconnection between art
history and scientific disciplines like comparative anatomy and physical anthropology,
particularly through a case study involving collaboration during the “Great War”.
Prown [18] had previously parsed the distinction in professions with the distinct role
of art historians being differentiated through a focus primarily on artworks. He
delineates “formal analysis” and “stylistic analysis”, emphasizing that these
concentrate on the art object’s configuration and style. Prown posits that style, as a
Forum for Art Studies 2024, 1(1), 393.
5
blend of form and distinctive manner, is culturally expressive, thus making it valuable
for understanding broader cultural contexts beyond art history.
In 2015, Locher et al. [19] explored the impact of formal art training on the
perception and aesthetic judgment of art compositions. Their study investigates
whether such training deepens the analysis of compositional elements, influencing
both the visual exploration and aesthetic evaluations of artworks, particularly those
that exhibit variations in balance and symmetry. Nelson [20] followed this in 2017,
focusing on the use of Visual Thinking Strategies (VTS) in art and information literacy
instruction. VTS, a method developed by Housen and Yenawine, utilizes art
discussions to encourage participation, critical thinking, and detailed observation.
Nelson illustrates how VTS not only enhances research strategies but also connects to
formal and critical analysis in art history. Collectively, these studies highlight the
dynamic nature of formal analysis in art history education, emphasizing its evolving
methodologies, pedagogical significance, and potential for cross-disciplinary
application and innovative teaching practices.
These examples highlight that teaching formal analysis and close looking in art
history classes involves guiding students to develop the skills of close observation,
critical thinking, and effective communication of their analyses. It is essential for
students to understand the elements of art and principles of design, as well as to learn
how to analyze and talk about art using formal analysis as a starting point. Teaching
formal analysis and close looking for art history classes is crucial for developing
student skills in analyzing and interpreting artworks. It involves close observation,
critical thinking, and understanding the elements of art and principles of design, which
are essential for gaining insights into the histories of art and effectively
communicating about art [21]. At the same time, scholarship on teaching and learning
associated with formal analysis in art history has shown a significant shift towards
enhancing student engagement and learning outcomes, especially in online
environments. Kutis [2] explores this evolution and presents a case study that
underscores the importance of scaffolding assignments in online courses, particularly
in introductory art history classes. This approach involves breaking down the standard
formal analysis assignment into smaller, manageable tasks using the learning
management system (LMS), Canvas.
The transition of college courses to online platforms presents unique challenges,
especially in maintaining effective faculty-student interaction and offering real-time
guidance for complex projects. Kutis highlights those scaffolding assignments can
significantly boost student success. This method entails creating smaller tasks that
progressively build toward a larger, more intricate assignment. Scaffolding not only
allows students repeated practice in developing their visual analysis and writing skills
but also provides opportunities for frequent instructional feedback. These practices are
recognized as best practices in the pedagogy of teaching and learning. By focusing on
specific aspects of the larger formal analysis task in each mini-assignment, students
can refine their skills step by step. This method also allows for continuous teacher
input, crucial in an online setting where direct interaction is limited. The approach
demonstrates quantitatively that such scaffolding improves student learning and
success, suggesting that this technique is beneficial for students in developing critical
skills necessary for formal analysis in art history. This strategy aligns well with the
Forum for Art Studies 2024, 1(1), 393.
6
current emphasis on self-regulated learning, catering to the evolving needs of students
in the digital age.
Transitioning from Kutis [2], who emphasizes the use of LMS to deconstruct
formal analysis into manageable steps, we enter the domain of text-to-image AI
technology, which offers a groundbreaking shift in art history education. While these
tools have been widely explored for artmaking purposes, their potential in art history
teaching remains largely untapped [22]. AI art prompts, which are textual instructions
guiding text-to-image generators to create original artworks, utilize AI algorithms
trained on extensive visual art datasets. The process allows the AI to mimic diverse
artistic styles and techniques. Therefore, advancing into the realm of Text-to-Image
artificial intelligence, significant developments, notably with Dall-E and Stable
Diffusion, present intriguing possibilities for visual arts education. Dehouche and
Dehouche [23] delve into the impact of these AI programs on art history, aesthetics,
and technical instruction. Analyzing a large sample of 72,980 Stable Diffusion
prompts, their research contributes to the formalization of this innovative art creation
method. Their findings indicate that Text-to-Image AI holds the potential to
revolutionize art education, offering cost-effective and inventive avenues for
experimentation and expression. Yet, these advancements also bring to light concerns
regarding the ownership of AI-generated artworks, signaling an urgent need for novel
legal and economic models to protect the rights of artists [24]. The shift towards AI in
art history education represents a significant step in exploring new methodologies and
addressing contemporary challenges in the field.
Following these studies, Gu et al. [12] provide a comprehensive overview of
prompt engineering—a technique involving the augmentation of large pre-trained
models with task-specific hints, known as prompts, to adapt the model to new tasks.
This paper surveys the latest research in prompt engineering across different types of
vision-language models like multimodal-to-text generation models (e.g., Flamingo),
image-text matching models (e.g., CLIP), and text-to-image generation models (e.g.,
Stable Diffusion). Gu et al. [12] discuss the summaries of these models, prompting
methods, applications, and associated responsibility and integrity issues. Their work
underscores the significance of prompt engineering in bridging vision-language
models with real-world tasks and the challenges and future directions in this rapidly
evolving field. Both these studies highlight the transformative potential of AI in visual
arts education, especially in teaching art history and formal analysis, while also
acknowledging the ethical and legal considerations that accompany this technological
leap.
In the context of visual arts education, integrating text-to-image AI raises
questions about the role of artistic intentionality. While there is a concern that relying
on AI systems might limit the range of expressions available to students, potentially
leading to a homogenization of artistic output, these technologies also offer unique
benefits [25]. I art generators can foster a deeper understanding and appreciation of
different artistic styles, movements, and methods among art history students. By
inputting specific prompts and choosing styles, students can generate captivating AI-
created art that reflects various historical periods and artistic approaches. In teaching
art history, the use of AI art generators can be particularly impactful. They provide a
novel, interactive way for students to engage with and analyze historical artworks,
Forum for Art Studies 2024, 1(1), 393.
7
styles, and techniques. This approach not only encourages a more active learning
process but also serves as a valuable tool for igniting creativity and imagination,
allowing students to explore the vast possibilities of artistic expression through a
modern, digital lens.
3. Methodology
The methodology for this mixed-methods study focused on examining the
efficacy of AI art generators in teaching art history. The study was conducted at a
private, four-year liberal arts institution located in the suburban area of St. Louis,
Missouri. Participants included undergraduate and graduate students enrolled in an
Ancient Art course, which covered periods from ancient Mesopotamia through the fall
of Rome. The course was offered online, presupposing a basic understanding of digital
tools among the students. The study sample comprised 24 undergraduate and graduate
students of which 8 were respondents from the College of Arts and Humanities all
majoring in Art History and Visual Culture. The primary goal of the project was to
assess pedagogical best practices for the use of AI art generators in art history
education, focusing on student perceptions, performance, and feedback, supplemented
with instructor observations.
Participants were tasked with a specific assignment (Table 1): to use an AI art
generator to recreate a piece of art from the course material. The students were
encouraged to use free AI art generators such as Midjourney, Stable Diffusion,
Craiyon, DALL·E 2, or, for those with access, ChatGPT-4 with DALL·E 3. They were
instructed to apply formal analysis techniques learned in the class to craft prompts that
would guide the AI in generating images as close as possible to the original artworks.
The students were not allowed to use identifiable descriptors such as the work’s title,
region, or style in their prompts.
The study utilized both qualitative and quantitative data collection methods. An
online survey, conducted at the beginning and end of the Fall 2024 term, collected
demographic information and students’ expectations and experiences with AI
generative art. The survey included open-ended questions allowing students to express
their views on the pedagogical effectiveness of AI in art history. Data from the surveys
were collected using Qualtrics to ensure privacy and anonymity. Descriptive statistics
were calculated from the survey responses for comparative analysis.
Additionally, the study analyzed the final AI-generated images and reflective
essays submitted by the students. These artifacts were evaluated to understand the
depth of students’ formal analysis skills and their ability to translate these skills into
AI prompts. Instructor feedback and observations provided further insights into the
learning outcomes and the pedagogical effectiveness of integrating AI art generators
into art history education. This mixed-methods approach aimed to provide a
comprehensive understanding of how AI art generators can enhance the teaching and
learning of art history, specifically focusing on the development and application of
formal analysis skills in a digital context.
Forum for Art Studies 2024, 1(1), 393.
8
Table 1. AI art history assignment.
Assignment instructions
In this assignment, you will engage with art history in an innovative way by combining your knowledge of art with cutting-edge
AI technology. You will select an artwork covered in our class and attempt to recreate it using an AI art generator. This exercise
will not only reinforce your understanding of the selected work but also expand your skill set in digital art creation and prompt
engineering.
Component
Description
Assignment name
AI art history assignment
Objective 1
Demonstrate understanding of a historical artwork through detailed formal analysis.
Objective 2
Apply prompt engineering skills to recreate the artwork using AI technology.
Objective 3
Reflect on the intersection of art history and AI in the creation and study of art.
Step 1: Select an artwork
Choose a piece of art covered in this class. It should be a work that resonates with you,
from any period discussed.
Step 2: AI art generators
Before creating your artwork, you must understand the capabilities and limitations of the
AI tool you’ll be using. Here are some options:
Midjourney (Web-based): Midjourney https://www.midjourney.com/home (note: must be
accessed via Discord account)
Stable diffusion (Web-based): Stable diffusion https://stability.ai/stable-diffusion/
Craiyon (formerly DALL·E mini, Web-based): Craiyon https://www.craiyon.com/
DALL·E 2 (Subscription-based): DALL·E 2 https://openai.com/dall-e-2
ChatGPT-4 with DALL·E 3 (Subscription-based): For those with access, use your
ChatGPT-4 interface. https://openai.com/blog/chatgpt
Step 3: Formal analysis
Write a detailed visual description of your chosen artwork. Focus on the formal elements
such as material, composition, subject matter, lighting, line, form, and structure. Avoid
using the title of the work; use only visual descriptors.
Step 4: Create AI artwork
Craft prompts using the formal analysis to guide the AI in recreating the artwork. This
may require multiple attempts, refining the prompt with each iteration.
Step 5: Reflective essay
Write a 750–1000 word essay discussing the process of translating art historical analysis
into AI prompts, challenges in the recreation process, insights about the artwork, and your
view on the role of AI in art creation and history.
Submission requirements
Submit the best AI-generated image and the reflective essay. Include screenshots or
descriptions of prompt iterations.
Grading rubric
Criteria
Excellent (90%–100%)
Good (70%–89%)
Satisfactory (50%–
69%)
Needs improvement
(<50%)
Formal analysis
In-depth and insightful
analysis.
Comprehensive analysis
with minor gaps.
Basic analysis with
some missing elements.
Superficial or
significantly
incomplete.
Prompt engineering
Highly effective
prompts; artwork
closely resembles the
original.
Effective prompts;
artwork resembles the
original with minor
differences.
Adequate prompts;
artwork somewhat
resembles the original.
Ineffective prompts;
artwork does not
resemble the original.
Reflective essay
Insightful reflection,
and excellent
understanding of AI’s
role in art.
Good reflection, a clear
understanding of AI’s
role.
Basic reflection, some
understanding of AI’s
role.
Limited or unclear
reflection, lacks
understanding of AI’s
role.
Creativity and iteration
process
Demonstrates high
creativity and
thoughtful iteration.
Good creativity and
adequate iteration.
Some creativity, but
limited iteration.
Lacks creativity and
effective iteration.
Overall presentation
Excellent organization
and clarity.
Good organization and
clarity.
Adequate organization
and clarity.
Poor organization and
lack of clarity.
Forum for Art Studies 2024, 1(1), 393.
9
4. Results
In the study assessing the integration of AI in art history education, a diverse set
of demographic data was collected through a survey instrument. The participant group
consisted of 8 students from a digital art history course, providing a range of insights
into their educational backgrounds and personal demographics. The majority of
participants were at an advanced academic level, with 62.5% (n = 5) identifying as
graduate students. This was complemented by an equal distribution among
sophomores, juniors, and seniors, each category comprising 12.5% (n = 1) of the
respondents. Such a distribution suggests a varied academic perspective within the
course, potentially influencing how students engage with and perceive the integration
of AI in their studies.
Regarding age, the participants predominantly fell within the 25–34 age bracket,
making up 62.5% (n = 5) of the responses. This was followed by the 18–24 age group,
which constituted 25% (n = 2) of the participants, and a smaller representation in the
35–44 age group at 12.5% (n = 1). This age range indicates a mature student body,
likely to bring a depth of life experience to their engagement with the course material.
The survey also revealed a homogenous gender identity among participants, with
100% (n = 8) identifying as female. This gender distribution could provide unique
perspectives in the interpretation and creation of art through AI tools.
In terms of ethnicity, the survey showed that none of the participants identified
as Hispanic/LatinX. The racial composition was predominantly White/Caucasian,
accounting for 70% (n = 7) of the respondents, with American Indian or Alaskan
Native and Black or African-American students representing 10% (n = 1) and 20% (n
= 2), respectively. This demographic makeup provides insights into the cultural
diversity of the participants and how it might impact their approach to and
understanding of art history in the context of AI. Additionally, the survey looked into
the educational background of the participants’ immediate families. A significant
majority, 75% (n = 6), reported that a family member completed an undergraduate
degree, while 12.5% (n = 1) had family members who completed some college credits
but did not finish their degree, and another 12.5% (n = 1) had family members with a
Master’s degree. This information is indicative of the participants’ educational
environments and the potential influence of their family backgrounds on their
academic pursuits.
Finally, the mode of class attendance was predominantly online, with 87.5% (n
= 7) of the participants primarily taking classes in this format, and only 12.5% (n = 1)
attending face-to-face classes. This data is particularly relevant, given the online
nature of the course and the use of digital tools like AI art generators. The preference
for online classes reflects the growing trend of digital learning environments in higher
education and provides context for how students interact with and perceive online
learning tools, including those used for creating AI-generated art.
In the study exploring the integration of AI art generators in an art history course,
student responses provided valuable insights into their experiences with these
emergent technologies. The survey responses revealed how students interacted with
AI art generators and their perceptions of the technology’s impact on their
understanding of art history.
Forum for Art Studies 2024, 1(1), 393.
10
A key aspect of the study was to ascertain students’ prior experience with AI art
generator tools. Out of the 8 participants, 37.5% (n = 3) reported having used AI art
generators before the class, while the majority, 62.5% (n = 5), had not. This data
indicates a relatively new exposure to AI art generation tools among most students,
suggesting that the course was likely their first encounter with applying these
technologies in an academic setting.
Regarding the specific AI art generators used for the assignment, the survey
showed varied usage among students. Craiyon was the most popular choice, used by
50% (n = 4) of the students. Midjourney and ChatGPT-4 with DALL·E 3 each were
chosen by 12.5% (n = 1) of the participants. Notably, 25% (n = 2) of the students opted
for other unspecified AI tools. This diversity in tool selection illustrates the range of
available AI art generators and students’ exploratory approach in selecting the tool
that best suited their needs for the assignment.
In terms of the challenge presented by using these tools, the responses varied. A
total of 75% (n = 6) of the students found using the AI art generator to be very easy or
somewhat easy, while 12.5% (n = 1) rated it as neutral and another 12.5% (n = 1) as
somewhat challenging. None of the students reported finding the use of AI art
generators very challenging. This suggests that, overall, students were able to navigate
the AI tools with relative ease, encountering minimal difficulties in their usage.
Finally, when asked about the extent to which the AI tool helped them understand
the artwork they were recreating, student responses were mixed. While 25% (n = 2)
reported that the AI tool did not help them at all, another 25% (n = 2) found it
moderately helpful, and 25% (n = 2) thought it was very helpful. An additional 12.5%
(n = 1) of the students felt the AI tool was extremely helpful, and 12.5% (n = 1) found
it only slightly helpful. This spread of responses indicates varied perceptions of the AI
tool’s effectiveness in enhancing their understanding of the artworks, reflecting the
subjective nature of how students engage with and benefit from technology in learning.
Students were then given a free-response question about difficulties encountered
while using the AI tool revealed a range of experiences among the students. Selected
key quotes illustrate the challenges faced. For instance, one student expressed a
significant challenge, stating, “It could not recreate the image because the image was
too specific and complex.” This comment highlights a limitation of AI tools in dealing
with intricate and highly detailed artworks, reflecting the difficulty in capturing the
essence of complex pieces. Another student observed the developmental stage of AI
platforms, noting, “I did not have a ton of difficulties. But it is more so that I think that
the AI platforms themselves are still being developed and though they may get close
to the look of a piece, they still don’t compare to the entire feeling a piece of art shows
or entirely captures its unique features.” This response points to the gap between AI-
generated images and the depth of emotion or unique characteristics inherent in
original artworks. The challenge of creating effective prompts was highlighted by a
student: “My biggest difficulty was wording the prompts correctly to get the outcome
I was looking for.” This difficulty underscores the importance of precise language in
guiding AI to produce desired results, a key aspect of prompt engineering. Another
student mentioned logistical issues, particularly regarding access to AI tools: “Mainly
finding a program that didn’t require payment. In creating the image, itself, limbs were
a big struggle and unnatural poses were very common.” This statement points to the
Forum for Art Studies 2024, 1(1), 393.
11
practical challenges of finding accessible AI tools and the technical limitations of AI
in accurately rendering human figures. These responses collectively indicate a variety
of challenges faced by students, ranging from the technical limitations of AI tools in
capturing complex artwork details to the intricacies involved in crafting effective
prompts and accessing suitable platforms.
The concluding part of the results section delves into students’ perceptions of the
role of AI tools in art history education, their views on how AI can contribute to the
study, and their overall feedback on the experience. Responses to whether AI tools
have a place in art history education were evenly split across the spectrum. Exactly
25% of the students strongly agreed while another 25% agreed, suggesting a favorable
view of AI integration in art history education. However, an equal proportion (25%)
were neutral, and 25% disagreed, indicating some skepticism or uncertainty about the
role of AI in this field. This diversity in opinions reflects the varying levels of
acceptance and readiness to embrace AI as a part of art historical studies.
Students’ insights on how AI could contribute to the study of art history revealed
diverse perspectives. One student noted the potential of AI in “aiding in image
recognition, restoration, cataloging, authenticity verification, personalized content
recommendation, contextual analysis, and immersive experiences,” underscoring AI’s
multifaceted utility in enhancing the accessibility and understanding of artistic
heritage. Another student expressed uncertainty, stating, “I don’t know enough about
it yet but I don’t see how it has an impact on historical artworks except to cause
confusion and misrepresentation.” This response highlights a concern about AI’s
potential to distort the understanding of historical art. Another perspective suggested
that AI’s growing popularity could offer new insights into both modern and ancient
art, including the recreation of lost artworks. Conversely, one student felt that while
AI helped in deepening their engagement with the artwork, it did not significantly
contribute to their overall understanding.
Finally, students’ feedback on their experience with the study was generally
positive but varied. One student enthusiastically shared, “I thoroughly enjoyed this
assignment and I would be very interested in doing another AI assignment in my other
courses!” Another remarked on the assignment’s ease: “crazy easy”. One student
appreciated the innovative nature of the task, finding it both challenging and fun.
However, a contrasting view was expressed by a student who felt the exercise was
interesting but did not significantly aid in learning or understanding art history. These
responses collectively suggest that while there is excitement and appreciation for the
innovative use of AI in art history education, there is also a degree of skepticism and
concern about its effectiveness and impact on learning. This range of sentiments
highlights the complex and nuanced nature of integrating emerging technologies like
AI into traditional academic disciplines.
5. Analysis
The split in student responses regarding the role of AI tools in art history
education indicates a field in transition. The fact that half of the students either agreed
or strongly agreed about the role of AI suggests an openness to embracing new
technologies in academic settings. This positive reception might be driven by the
Forum for Art Studies 2024, 1(1), 393.
12
potential of AI to offer novel ways of interacting with and interpreting art. However,
the presence of neutrality and disagreement in equal measure also points to
reservations and uncertainties. These could stem from concerns about the efficacy of
AI in capturing the essence of artworks, its impact on traditional learning methods, or
apprehension about the rapid pace of technological change in educational settings.
The varied opinions on AI’s contributions to art history reflect a spectrum of
expectations and understandings of the technology. The recognition of AI’s potential
in tasks such as image recognition, restoration, and contextual analysis suggests an
acknowledgment of its utility as a tool for enhancing certain aspects of art historical
study. However, skepticism about AI’s impact on the interpretation of historical
artworks, and concerns about misrepresentation, highlight critical reservations. This
dichotomy underscores the need for a balanced approach in integrating AI into art
history education, one that leverages its strengths while being mindful of its limitations.
The generally positive feedback on the AI assignment, with some students
expressing enthusiasm and others noting its ease of use, suggests that such innovative
approaches can enhance student engagement and interest. The appreciation for the
assignment’s creative and challenging aspects indicates that AI can add value to the
learning process by providing new experiences and perspectives. However, the
viewpoint that the assignment did not significantly aid in learning or understanding
points to a gap between the novelty of the technology and its pedagogical effectiveness.
This suggests that while AI can serve as an engaging tool, its integration into the
curriculum needs to be carefully structured to ensure that it complements and enhances
traditional learning objectives.
The analysis of the survey responses indicates that while there is enthusiasm and
potential for integrating AI in art history education, there is also a need for careful
consideration of how it is implemented. The effectiveness of AI in enhancing learning
outcomes in art history lies not just in its technological capabilities, but in how it is
integrated into the curriculum and aligned with pedagogical goals. The diverse student
experiences and perceptions highlight the importance of addressing concerns about
AI’s limitations and ensuring that its use supports rather than supplants traditional
methods of teaching and learning in art history.
Perhaps one of the most successful came from a student who voiced the most
concern over how these tools would be used in their reflective essay. In discussing her
generation of the Parthenon, the student wrote:
I think AI in general is a really interesting and useful resource, but it depends on
how people use it and why. When people use AI as a tool for inspiration, fostering
creativity, or enhancing personal artistic endeavors, it can serve as a helpful
resource to help those who are struggling. The way that AI can generate ideas or
assist people in creating captivating pieces can come in handy for artists seeking
new avenues of expression. On the other hand, if people use text or art AI
generators to either pass off the products as their own or to purposely steal work
from others, that is not right. For instance, one of the biggest issues with AI art is
the fact that a majority of AI art generators use stolen art made by real artists
without their permission.
Forum for Art Studies 2024, 1(1), 393.
13
At the same time, the student followed the assignment as outlined and used
several iterative prompt strategies in order to arrive at the image. These prompts
included:
The analysis of student artifacts, including the images generated by AI tools and
their reflective essays, provides crucial insights into the effectiveness of AI in
facilitating the learning of art history and formal analysis. One example, the Nike of
Samothrace, was described as a “Marble woman with no head and no arms, right leg
stepping forward, clothed in a dress that sticks to her body with wings opened for
flight” (Figure 1). The student noted that free tools like Craiyon had issues with
anatomical accuracy, resulting in limbs with mistakes and awkward poses. This
example underscores the limitations of AI tools in accurately capturing complex and
nuanced details of human anatomy and poses, a crucial aspect of formal analysis in art
history.
Figure 1. Student image replicating Nike of Samothrace.
Another issue was observed in instances where students did not strictly adhere to
the assignment guidelines about avoiding specific periods, cultures, or style in the
prompt. An example of this was the use of a prompt describing an “ancient Roman
statue of a dying Celtic warrior on the ground, pushing himself up with one arm”
(Figure 2). The generator failed to produce the desired results, as most images did not
Forum for Art Studies 2024, 1(1), 393.
14
depict the figure in a prostrate position, and the prompts were too specific, limiting the
AI’s interpretative ability.
Figure 2. Student image replicating Dying Gaul.
Conversely, some students achieved success with their AI-generated images. A
notable example involved the Laocoon (Figure 3), where a student produced six
variations using Stable Diffusion. The prompts used were detailed yet open-ended,
such as describing a “muscular, middle-aged man with short, tousled hair, being
attacked by a serpent.” The use of detailed descriptors led to images resembling aged
engravings, showing the potential of AI in capturing the essence of art with proper
prompting.
Forum for Art Studies 2024, 1(1), 393.
15
Figure 3. Student image replicating Laocoon.
One of the most successful examples was from a student who initially expressed
concerns over AI usage in art. In their reflective essay, the student acknowledged AI’s
potential as a source of inspiration and creativity but cautioned against its misuse. As
the student wrote:
I think AI in general is a really interesting and useful resource, but it depends on
how people use it and why. When people use AI as a tool for inspiration, fostering
creativity, or enhancing personal artistic endeavors, it can serve as a helpful
resource to help those who are struggling. The way that AI can generate ideas or
assist people in creating captivating pieces can come in handy for artists seeking
new avenues of expression. On the other hand, if people use text or art AI
generators to either pass off the products as their own or to purposely steal work
from others, that is not right. For instance, one of the biggest issues with AI art is
the fact that a majority of AI art generators use stolen art made by real artists
without their permission.
At the same time, their output was exceptional. This student meticulously
followed the assignment, using iterative prompts to describe the Parthenon (Figure 4)
with varying degrees of detail, such as:
• A grand peripteral, Doric order temple made of white marble. 8 columns on the
front and 12 columns on the sides. It is missing the roof and tiles, and the marble
is weathered from age.
Forum for Art Studies 2024, 1(1), 393.
16
• A grand peripteral, Doric-order temple made of white marble that has now faded
with age. the temple has 8 columns on the front and 12 columns on the sides. The
roof of the temple is missing and the pediments are severely damaged.
• A massive peripteral temple made of yellowed marble with 8 Doric order
columns on the front and 12 Doric order columns on the sides. The roof of the
temple is missing and the pediments are severely damaged.
• A massive peripteral temple made of yellowed marble with 8 Doric order
columns on the front and 12 Doric order columns on the sides. The roof and
interior of the temple is destroyed and the pediments are severely damaged.
• A massive peripteral temple made of yellowed marble with 8 Doric order
columns on the front and 12 Doric order columns on the sides. The roof of the
temple is destroyed and the pediments are severely damaged. The interior of the
temple was destroyed, leaving the interior empty and open.
• A long peripteral temple made of weathered, yellowed marble with 8 plain Doric
order columns on the front and 12 plain Doric order columns on the sides. The
temple’s roof is missing and the pediments are severely damaged and broken.
The interior of the temple is gone, leaving the inside empty and open.
• A long peripteral temple made of weathered, yellowed marble with 8 plain Doric
order columns on the front and 12 plain Doric order columns on the sides. The
temple’s roof is missing and the pediments are severely damaged and broken.
The interior of the temple is gone, leaving the inside empty and open. It sits on
rocks in front of a blue, cloudy sky.
Figure 4. Student image replicating the Parthenon.
Forum for Art Studies 2024, 1(1), 393.
17
The gradual evolution of these prompts resulted in a nuanced representation of
the Parthenon, exemplifying the effective use of AI in understanding and recreating
historical architecture.
The student artifacts and reflective essays from this study highlight both the
challenges and possibilities of using AI in art history education. While AI tools can
sometimes struggle with complex imagery and nuances, with thoughtful and detailed
prompting, they can also provide valuable insights and creative interpretations of
historical artworks. The success of these AI-generated images in educational settings
is contingent on the precise and informed use of language in prompts, coupled with a
deep understanding of the artworks being described.
6. Recommendations
In light of the findings and analysis of the study on integrating AI tools in art
history education, several recommendations can be made to enhance the learning
experience and outcomes. Firstly, it is advisable to encourage the use of more
advanced and sophisticated AI tools such as Stable Diffusion and Dalle-3 in art history
courses. These tools have shown a higher capability in accurately interpreting and
rendering complex art-related prompts. Their advanced algorithms can better handle
the intricacies and nuances of artworks, making them more suitable for educational
purposes in art history. By utilizing these more powerful tools, students can achieve a
closer approximation to the original artworks in their AI-generated images, thereby
enhancing their understanding of formal analysis and artistic techniques.
Secondly, it is crucial to ensure that students have a solid understanding of formal
analysis and the objectives of the assignment. This understanding is key to effectively
using AI art generators. Students should be guided to focus on describing the formal
elements of artworks, such as line, shape, color, texture, and composition, without
including descriptors that pertain to the period, culture, or style. Clear instructions and
examples should be provided to help students craft prompts that are focused solely on
visual and formal aspects. This focus will aid in reducing misinterpretations by the AI
and lead to more accurate recreations of the artworks.
Additionally, offering training and guidance in the art of prompt engineering is
recommended. Prompt engineering is a critical skill in the effective use of AI tools, as
the quality of the output heavily depends on the input prompt. Workshops or
instructional sessions can be conducted to teach students how to construct effective
and precise prompts. These sessions could include exercises in translating visual
analysis into written descriptions and iteratively refining prompts based on AI-
generated outputs. By enhancing their prompt engineering skills, students can better
navigate the capabilities and limitations of AI tools, leading to more successful
outcomes in their assignments.
Finally, establishing robust monitoring and feedback mechanisms is important.
Continuous monitoring of how students interact with and utilize AI tools can provide
valuable insights into their learning process. Regular feedback from instructors will
help students fine-tune their approach to using AI in art history. This feedback could
take the form of critiques of the AI-generated images and discussions of the reflective
essays. Through this iterative process of creation, feedback, and refinement, students
Forum for Art Studies 2024, 1(1), 393.
18
can progressively improve their skills in formal analysis and their understanding of
how AI can be used as a tool in art history education. Implementing these
recommendations can significantly enhance the efficacy and educational value of AI
tools in art history courses, ensuring that they serve as a complement to traditional
learning methods and contribute positively to students’ understanding of art and its
analysis.
7. Conclusion
This study has provided valuable insights into the potential and challenges of
integrating AI tools into art history education. The key takeaways from the research
underscore both the opportunities and the nuanced complexities involved in using AI
for teaching and learning in the field of art history. The study revealed that while AI
tools, particularly advanced ones like Stable Diffusion and Dalle-3, have considerable
potential in aiding formal analysis and close looking at art history, their effectiveness
is heavily dependent on how they are utilized. Students’ varied responses and
experiences highlighted the importance of precise prompt engineering and a deep
understanding of formal analysis. The challenges faced, such as difficulties in
accurately capturing complex artworks and issues with overly specific prompts, point
to the limitations of current AI technologies in fully grasping the nuances of artistic
expression. The research demonstrated that AI tools could significantly contribute to
art history education, particularly in enabling students to engage with artworks in
innovative ways. These tools offer novel means of visualizing and interpreting art,
providing students with alternative perspectives and methods to analyze and
understand artworks. When used effectively, AI can enhance students’ creativity and
analytical skills, offering a complementary approach to traditional art history
methodologies.
Looking ahead, there are several areas for future research that could further
elucidate the role of AI in art history education. One key area is the development of
more sophisticated AI algorithms that can better handle the complexities of art
interpretation. Research could also explore the integration of AI tools in other aspects
of art history education, such as in the study of iconography, art criticism, and art
theory. Additionally, longitudinal studies assessing the long-term impact of AI tool
usage on students’ learning outcomes and understanding of art history would be
valuable. Furthermore, exploring interdisciplinary approaches that combine AI with
other technological advancements, such as virtual reality, could open up new frontiers
in art history education. While AI tools present exciting opportunities for enhancing
art history education, their integration requires careful consideration and strategic
planning. Future research and development in this field should focus on overcoming
the current limitations of AI technologies, enhancing their pedagogical application,
and exploring innovative ways to blend these tools with traditional art historical
methods. The ultimate goal should be to leverage AI not as a replacement, but as an
enriching complement to the established practices of art history teaching and learning.
Conflicts of interest: The author declares no conflict of interest.
Forum for Art Studies 2024, 1(1), 393.
19
References
1. Donahue-Wallace K, La Follette L, Pappas A. Teaching art history with new technologies: Reflections and case studies.
Cambridge Scholars Publishing; 2009.
2. Kutis B. Scaffolding the Formal Analysis Assignment in Art History Courses to Promote Learning. Journal of Teaching and
Learning with Technology, 2020; 9(1).
3. Dow AW. Composition: A series of exercises in art structure for the use of students and teachers. University of California
Press; 2023.
4. Black CV, Barringer T. Decolonizing Art and Empire. The Art Bulletin, 2022; 104(1), 6-20.
5. Graham ME, Bailey C. Digital images and art historians–Compare and contrast revisited. Art Libraries Journal, 2006; 31(3),
21-24.
6. Alpers P. Ut Pictura Noesis? Criticism in literary studies and art history. In: New Directions in Literary History. Routledge;
2022. pp. 199-220.
7. Hutson J, Olsen T. Digital humanities and virtual reality: A review of theories and best practices for art history. International
Journal of Technology in Education (IJTE), 2021; 4(3), 491-500.
8. Hutson J, Olsen T. Virtual Reality and Art History: A Case Study of Digital Humanities and Immersive Learning
Environments. Journal of Higher Education Theory and Practice. 2022; 22(2), 49-64. doi: 10.33423/jhetp.v22i2
9. Zullich M, Macovaz V, Pinna G, Pellegrino FA. An artificial intelligence system for automatic recognition of punches in
fourteenth-century panel painting. IEEE Access, 2023; 11, 5864-5883.
10. Hutson J, Robertson B. Exploring the Educational Potential of AI Generative Art in 3D Design Fundamentals: A Case Study
on Prompt Engineering and Creative Workflows. Global Journal of HUMAN-SOCIAL SCIENCE: A Arts & Humanities-
Psychology, 2023; 23(2).
11. Gu J, Han Z, Chen S, et al. A systematic survey of prompt engineering on vision-language foundation models. arXiv, 2023;
arXiv:2307.12980.
12. Walton N. There are no formal elements. Debates in Art and Design Education, 2020; 66.
13. Gasper-Hulvat M. Active learning in art history: A review of formal literature. Art History Pedagogy & Practice, 2017; 2(1),
2.
14. Avci H, Pedersen S, Thomas A. Writing a Formal Analysis of Art in a Game-Based Learning Environment. In: EdMedia+
Innovate Learning. Association for the Advancement of Computing in Education (AACE); 2020. pp. 669-671.
15. Elgammal A, Liu B, Kim D, et al. The shape of art history in the eyes of the machine. In: Proceedings of the AAAI
Conference on Artificial Intelligence. 2018.
16. Freedman K. Teaching visual culture: Curriculum, aesthetics, and the social life of art. Teachers College Press; 2003.
17. Olin M. Formal Analysis: Art and Anthropology. Ideas of’Race’in the History of the Humanities, 2017; 89-111.
18. Prown JD. Style as evidence. Winterthur Portfolio, 1980; 15(3), 197-210.
19. Locher P, Krupinski E, Schaefer A. Art and authenticity: Behavioral and eye-movement analyses. Psychology of Aesthetics,
Creativity, and the Arts, 2015; 9(4), 356.
20. Nelson A. Visual Thinking Strategies from the museum to the library: Using VTS and art in information literacy instruction.
Art Documentation: Journal of the Art Libraries Society of North America, 2017; 36(2), 281-292.
21. Dolšina M. Educational role of art history as a school subject area in programmes of formal education in Slovenia. The
Journal of Education, Culture, and Society, 2015; 6(1), 227-241.
22. Mazzone M, Elgammal A. Art, creativity, and the potential of artificial intelligence. In: Arts. MDPI; 2019.
23. Dehouche N, Dehouche K. What’s in a text-to-image prompt? The potential of stable diffusion in visual arts education.
Heliyon. 2023.
24. Jiang HH, Brown L, Cheng J, et al. AI Art and its Impact on Artists. In: Proceedings of the 2023 AAAI/ACM Conference on
AI, Ethics, and Society. 2023. pp. 363-374.
25. Hutson J, Harper-Nichols M. Generative AI and Algorithmic Art: Disrupting the Framing of Meaning and Rethinking the
Subject-Object Dilemma. Global Journal of Computer Science and Technology: D, 2023; 23(1).
