Leveraging Artificial Intelligence Tools in Digital Media Development

to Lower Development Barriers

Gejian Zhu

Department of Design, Goldsmiths, University of London, London, U.K.

Keywords: Artificial Intelligence, Digital Media Interaction, Low-Code Tools, Creative Technology, User

Experimentation.

Abstract: With the widespread adoption of artificial intelligence (AI) tools in digital media creation, whether AI tools

have lowered the technical barrier for non-professionals is worth questioning. Taking open source AI

platforms ml5.js and Ready Player Me as examples, this study examines whether non-programming users can

independently achieve interactive media effects, such as gesture recognition and virtual character control,

without conventional programming skills. Experimental observations on two design students were used to

assess task completion, perceived usability of the tools, and understanding of system logic. The experimental

results show that although participants finished the task successfully, their knowledge of the internal operation

of the tool was comparatively more limited. This study suggests that the AI platform significantly lowers the

threshold for development at the operational level but does not fully compensate for the gap in users'

perception of the system. This study provides a practical reference for creative technology education and AI

tool development and points out the possible risk of "black box usage".

1 INTRODUCTION

The rapid development of Artificial Intelligence

technology has resulted in a growing use of AI tools

in the digital media industry, especially in key

technical fields like image synthesis, speech

recognition, gesture tracking, and motion capture. AI

is increasingly replacing the traditional manual

modeling approach and programming development

process. In recent years, there has been a large

number of open-source or low-threshold AI

platforms, like ml5.js, MediaPipe, and Ready Player

Me, that have made interactive media content, which

previously required specialist programming skills,

available to people without technical backgrounds

(Billinghurst et al., 2015). This changing trend raises

a fundamental question: has AI technology truly

lowered the threshold for developing interactive

digital media technologies?

The creation of traditional interactive digital

media, especially with regard to complex

functionalities like mixed reality (MR), gesture

recognition, or the manipulation of virtual characters,

is often reliant on advanced computer vision

algorithms, sensor data processing, and skills in

object-oriented programming. For non-professionals,

the threshold to building such systems is usually high.

However, the development of artificial intelligence

technologies has somewhat changed this situation.

For example, ml5.js provides a modular gesture

recognition model, allowing for image input and

recognition processes to be completed with a few

lines of code; similarly, Ready Player Me provides for

the quick generation of 3D virtual characters from

photos, making it possible to integrate them into

interactive environments, thereby empowering "non-

programmers" to create interactive experiences.

These technologies make it possible for "non-

programmers" to participate in the creation of

interactive experiences.

However, systematic research to confirm the

competence and experience of non-professional users

in creating digital media interactions with the

assistance of AI tools remains scarce. The objective

of this research is to examine whether AI technology

has significantly reduced the threshold of digital

media technology development through empirical

experiments based on the observation of the process

and feedback of participants with non-programming

backgrounds in utilizing AI tools (e.g., ml5.js and

Ready Player Me) to accomplish basic interactive

tasks.

556

Zhu, G.

Leveraging Artiﬁcial Intelligence Tools in Digital Media Development to Lower Development Barriers.

DOI: 10.5220/0014362800004718

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 2nd International Conference on Engineering Management, Information Technology and Intelligence (EMITI 2025), pages 556-564

ISBN: 978-989-758-792-4

This research will aim to respond to the following

inquiries:

1. Can nonspecialist users utilize AI tools to

accomplish interaction tasks within a fairly short

time?

2. Do they have to depend on programming skills

while using?

3. What are the implications of the usability and

cognitive barriers present in AI tools for users from

varied backgrounds?

The importance of this research is not merely to

learn about the influence of AI on digital media's

technical threshold, but also to offer a practical

foundation for digital media education and tool

design in the future.

2 DEVELOPMENT STATUS OF

DIGITAL MEDIA

TECHNOLOGY

2.1 Developmental Threshold Levels

and Technological Challenges

Involving Digital Media Interaction

Technologies

With the passage of time, the development of

interactive technologies involving digital media like

virtual character animation, gesture recognition, and

mixed reality (MR) has called for the development of

advanced-level algorithms and complex system

architectures, and hence posed a serious technical

barrier to innovation. During previous eras, for

example, systems for gesture recognition normally

relied on tools like OpenCV, Kinect, or depth-sensing

cameras, together with computer vision approaches to

image processing, machine learning, and feature

abstraction, as a bid to create recognition systems

(Wachs et al., 2011). The development of MR

systems involves a form of programming and

platform technologies, ranging from Unity, C#, and

spatial map algorithms, as demanded by the

programmers (Billinghurst et al., 2015). Such a high-

level technical barrier has, on numerous occasions,

locked out nontechnical persons from the production

environment of interactive media content, leading to

a situation where content producers are mostly

engineers or development groups, as opposed to

mostly individual artists or designers.

2.2 AI-Powered Platform Development

and the Movement Towards

Low-Code Development

During the last decade, progress made on Artificial

Intelligence (AI) has gone a long way towards

fuelling the growth of "low-code" and "zero-code"

projects on digital media production. Several

platforms on AI have revolutionized conventional

methods, thereby reducing costs associated with

learning the abilities needed to construct complex

systems. For one:

The ml5.js library includes a range of pre-trained

models, including image classification, gesture

classification, and audio classification, that can be

easily used on the web using JavaScript in

applications that do not call for engineering skills,

including educational and interactive art solutions

(McCarthy, L., 2019).

MediaPipe thus supports end-to-end computer

vision solutions developed by Google that include

gesture recognition, face tracking, and detection of

human skeletal features, and is available on web and

mobile platforms (Lugaresi et al., 2019).

Virtual characters' platforms, such as Ready

Player Me, combine technologies such as three-

dimensional modeling, motion capture, animation

binding, and export procedures. Such combinations

allow a user to create interactive characters that can

be used on different platforms without needing to use

Blender or Maya (Wolf3D, 2021).

In recent years, low-code and no-code platforms

have emerged as a significant trend in software and

digital media development, substantially enhancing

the ability of “non-experts” to build applications

independently. Yan notes that such platforms enable

non-technical users to rapidly construct required

systems through visual components and drag-and-

drop logic, thus shortening development cycles and

reducing technical barriers (Yan, Z., 2021). Similarly,

Upadhyaya demonstrates that low-code tools play a

critical role in accelerating innovation, enabling faster

delivery, and supporting in-house solution

development in small and medium-sized enterprises

(Upadhyaya, 2023). These findings reinforce the

positioning of AI platforms in this study as “creator-

friendly tools” and provide a theoretical foundation

for their application in the field of digital media

interaction.

The arrival of these tools heralds a blurring of the

boundary between "technology-creation" and digital

media development, which is no longer exclusively

the domain of programmers, but has become a

creator-oriented platform design.

Leveraging Artiﬁcial Intelligence Tools in Digital Media Development to Lower Development Barriers

557

2.3 The Extent to Which AI Empowers

Nontechnical Creators: Synthesis of

Research

Despite the claim that tools of artificial intelligence

are actually made to reduce development barriers,

researchers are divided on the issue of empowerment

of the "non-professionals." Certain researchers

believe that AI tools increase the autonomy of

designers and foster interdisciplinary approaches

(Manovich, 2020); e.g., tools allow artists to combine

pictures and transform styles without the requirement

for programming knowledge, as shown through

applications such as RunwayML or DALL-E.

However, other researchers refute this notion,

declaring that systems of artificial intelligence can

cause a new form of a "black-box problem," where,

on the one hand, the models are easy to use but, on

the other, the essential mechanisms and logic on

which the outputs are based remain inaccessible to the

users, thereby limiting the system control exercised

by them (Burrell, 2016).

Additionally, educational technology experts

examined studies on the future use of artificial

intelligence when teaching digital media. More

specifically, Lee and Ko (2022) recognized that

integrating AI tools in interaction design classes can

significantly contribute to the value of student

outputs; nonetheless, a number of the students found

themselves challenged when altering the parameters

of the model and setting up file paths (Lee & Ko,

2022). As a result, although usability is improved

through the use of AI tools, a gap is observed between

usability and comprehensibility.

Previous literature suggests a movement from a

traditionally "expertise-oriented" development style

to a "creation-friendly" one aided by tools of artificial

intelligence, but empirical evidence is needed to

validate significant impacts on non-proficient

audiences. More particularly, the issue of whether or

not platforms based on AI allow for "de-

technicalized" development in domains where the

barrier to entry is very high, as is the case for gesture

recognition and mixed reality interaction, has yet to

be adequately examined based on actual behavior and

attitudes towards use. The experimental portion of

this work takes this as its starting point to assess to

what degree platforms based on AI allow non-

proficient end-users to create interactive media.

3 EXPERIMENTAL DESIGN

3.1 Purpose of the Experiment

The aim of the current study is to evaluate non-

experts' skill in making efficient use of core digital

media interaction functionalities through empirical

examinations of tools aided by artificial intelligence,

including ml5.js and Ready Player Me. The empirical

examinations are carried out to find out whether the

current AI tools have actually mitigated the problem

faced during the practical use of digital media

technologies. The core areas of interest include:

completion of tasks, understanding of the tools,

reliance on knowledge from previous programming,

and individual intuitions and perceptions met during

use.

3.2 Experiment Participants

This paper recruited two participants without any

prior experience in program development, who were

both digital media design and interaction design

students. They had a bit of creative knowledge but no

formal training in developing programs

systematically. Before the experiment started, all

participants filled out a background form to check on

their programming knowledge, experience with the

use of AI tools, and knowledge of digital media

interaction principles. The whole experiment was

video-recorded, and oral permission was obtained

from all participants.

3.3 Tool Platform Description

So that the experimental study is carried out on a

sustainable scale, two well-known open-source

platforms concerning artificial intelligence were

selected:

ml5. A client-side machine learning library meant

for progressive developers, featuring pre-trained

models that can be used for functions like gesture

recognition and image classification. The library

itself can be called from HTML and JavaScript.

Invocation by these technologies is simple, making it

suitable for the creation of interactive demos in a

browser environment.

Prepared Player Me. A cross-platform virtual

character generation tool that supports custom avatar

and animation binding. Users can upload selfies to

generate 3D character models and can realize simple

animation control in Unity and other platforms

through the web interface, which is suitable for MR

and virtual interactive scene construction.

EMITI 2025 - International Conference on Engineering Management, Information Technology and Intelligence

558

3.4 Experimental Task Design

Members are obliged to engage in the following two

activities:

Task 1 (based on ml5.js): Create a gesture

recognition interaction prototype aimed at starting or

stopping video playback by waving the hand.

Task 1 (using ml5.js): create a prototype that is

operable via hand movement interaction- i.e.,

"waving a hand to play/stop video playback."

Task 2 from Ready Player Me: Create a digital

avatar that can modify its behaviors, like waving or

nodding, on a web page based on users' interactions,

as when a button is clicked.

Character generation, loading, and the associated

logic that drives actions are illustrated and

incorporated via a friendly web interface.

In order to reduce information interference,

participants only received absolutely necessary

training materials and legitimate platform links; no

encrypted templates were provided, and participants

were expected to figure out the rationale themselves

for making use of the tools.

3.5 Methods for Collecting and

Assessing Data

Table 1 shows the empirical information will be

collected and analyzed using the following

methodologies

Table 1: Empirical Data Collection and Analysis Methods.

Data Collection

Tools

Recorded/Measured Content Instructions Example Questions/Variables

Measurement

Methods

Observation

Chart Sheet

Time to complete tasks -

Problems encountered - Help

needed - Frequency of

debu

incidents

Researchers

recorded the

experiment in

real time.

"Task 1 took 35 minutes" "I

encountered a model loading

error"

Process

observation +

record sheet

Participant

Survey (Post-

experiment

Questionnaire)

Using a five-point Likert

scale (1 = strongly disagree, 5

= strongly agree), measures:

1. Tool effectiveness 2.

Understanding of operating

procedures 3. Necessity of

programming 4. Perceived

creative freedom 5. Overall

satisfaction with AI tools

Quantitatively

measured

participants'

subjective

feelings.

"I was able to complete the

task successfully" "I

understood how the

interactive features work"

Questionnaire

(5-point Likert

scale)

Semi-Structured

Interviews

Deepen understanding of user

experience and perception

The recordings

were

transcribed and

coded for

analysis.

Which components were the

most difficult? Which areas

were the most challenging?

What was the main problem

that was effectively solved?

Have you noticed a decrease

in the technical barrier to

entry? Are you willing to

continue using AI tools in the

future?

Face-to-

face/online

interviews +

qualitative

analysis

A mixed-methods design will be utilized,

combining both quantitative and qualitative data

analysis methods.

Quantitative evaluation:

Compare the data for task completion time and the

self-assessment scale for the two participants.

Recognize the key aspects, such as the assignment

fulfillment, use of the code, and the level of

independence observed when carrying out the task.

Qualitative analysis:

Coding interviews aim to detect essential issues,

misconceptions concerning thought processes, and

positive assessments.

Assessing the validity of people's beliefs related

to the claim that "AI diminishes barriers to progress."

Its purpose is to assess the performance of AI-

assisted tools for first-time programmers through the

analysis of participants' actions and self-assessment.

Leveraging Artiﬁcial Intelligence Tools in Digital Media Development to Lower Development Barriers

559

4 ANALYSIS OF EMPIRICAL

RESULTS (FINDINGS AND

RATIONALE)

4.1 Brief Description of the

Experimental Methodology

This paper recruited two participants with no formal

programming backgrounds (hereafter referred to as

P1 and P2), one of whom was studying digital media

and the other graphic design; neither of the

participants had received any formal programming

training. After a questionnaire was used to collect

basic background information, the participants each

went through a step-by-step procedure involving two

experimental tasks:

First is using ml5.js to enable gesture recognition

for the control of video playback on a web page;

Second is creating virtual characters using Ready

Player Me and controlling their actions with web

buttons.

4.2 User Task Completion

Table 2 shows the task Completion and Support

Requirements of Participants. Observations revealed

that both participants completed the tasks on

assignment within the duration; however, Participant

2 had rational challenges (like setting up the resource

file path) getting through to Ready Player Me, but

managed to complete the task with a gentle push from

the researcher.

4.3 Analysis of Questionnaire Results

As shown in Table 3, the average score trend graph

illustrates that:

Has a high overall satisfaction and ease of use

rating (mean value 4.4), indicating that AI tools are

effective in reducing the burden of operation.

Has the lowest rating for "understanding of

system principles" (mean value 2.5), reflecting that

the black-box phenomenon of "only using but not

understanding" is common among non-professional

users. The lowest rating for "understanding of system

principles" (mean value 2.5) reflects that non-

professional users generally have the black box

phenomenon of "only using but not understanding".

Table 2: Task Completion and Support Requirements of Participants

Indicator P1 P2

Did Task 1 (Gesture Recognition) independently completed Independently completed

Did Task 2 (Virtual Characters) independently completed Required some technical assistance

Time spent on Task 1 (min) 35 min 45 min

Time spent on Task 2 (min) 40 min 60 min

Need to write code

(call predefined APIs + modify

parameters)

(Try to copy the code and debug the

structure.)

Need Documentation or guidance

required

Documentation consulted 1 time

Multiple times and in consultation with the

researcher

Table 3: The following are the Likert scale scores (1-5, higher means higher agreement)

Questionnaire items P1 P2

Q1: I was able to complete the main functions of the system independently 5 4

Q2: The difficulty of the task was within my understanding 4 3

Q3: I basically didn't need to write any code 5 4

Q4: The tool interface is user-friendly and quick to get started 4 4

Q5: It's easier than the traditional way 5 5

Q6: I understood how the interactive functions work 3 2

Q7: I was satisfied with the results 4 4

Q8: I'm willing to continue to use it in the future 5 4

Q9: I think that the AI tool lowered the technical threshold 5 4

Q10: I am uncomfortable with the “black box” mechanism 2 3

EMITI 2025 - International Conference on Engineering Management, Information Technology and Intelligence

560

Table 4: The following key findings were compiled from the semi-structured interviews.

ics Excer

ts from user feedbac

Ease of use of the tool "It's like

uttin

ether buildin

blocks, as lon

ou know which

ieces work."

Main difficulties "I'm stuck not knowing how to get the model to load."The Documentation is too

long, I don't want to read it."

Need to write code "Although I copied a few lines of code, I didn't write it; I just followed the tutorial

and changed the path."

Sense of control and

understandin

"The functionality works, but I don't quite know why it works."

Willin

ness to use in the future "Ver

ractical! I'd like to use it a

ain in a class

ect later."

4.4 Generalization of Interviews

Table 4 shows the following key findings were

compiled from the semi-structured interviews.

Overall, the interviews showed that AI tools have

good “task accessibility,” but users generally view

them as ‘toolboxes’ rather than “development

platforms,” and lack a deep understanding of model

parameters and system logic. However, users

generally viewed it as a “toolbox” rather than a

“development platform” and lacked a deep

understanding of the model parameters and system

logic

4.5 Testing

Research findings show that those who do not hold

specialized knowledge can use tools powered by AI

without human intervention to perform simple tasks

involving interaction. Additionally, resolutions can

be achieved for simple problems by reading

Documentation or collecting information from others.

Artificial intelligence tools have significantly

reduced the accessibility hurdles of coding and

system development, especially in the areas of

gesture recognition and virtual character creation.

However, the participants showed limited

awareness of the basic principles and logic governing

interactions, which is characteristic of an orthodox

"black-box use" practice.

AI tools perform well in the aspect of “assisted

creation”, but the participants' understanding of the

interaction logic and technical principles is still weak;

AI tools perform well in the aspect of “assisted

creation”, but the participants' understanding of the

interaction logic and technical principles is still weak.

AI tools perform well in “assisting creation”, but

there is still room for improvement in “technical

education”.

5 DISCUSSION AND

IMPLICATIONS (DISCUSSION

AND IMPLICATIONS)

5.1 Whether AI Tools Truly Lower the

Threshold of Digital Media

Development

Experiment outcomes show that users who had no

prior experience with programming were able to

successfully run virtual character control and gesture

recognition tasks within a short period. This result

suggests that artificial intelligence libraries like

ml5.js and Ready Player Me have meaningfully

reduced the barriers to the use of digital media

interaction technologies. In the past, acting out

similar functions required a high level of expertise in

image recognition algorithms, 3D modeling, and

script programming; however, in this case,

participants could carry out basic functionalities

using built-in APIs and graphical interfaces.

However, empirical data and findings from

interviews revealed one other dimension: while

participants cited greater convenience when using the

tool, they mostly lacked understanding from the

ground up of the principles regulating its function,

and a lack of information on the motivation behind

the inclusion of certain functionalities. Thus, this

suggests that tools driven by artificial intelligence

mainly reduce the 'barrier to creating superficial

results,' but the 'barrier to understanding and learning

the technology' endures.

5.2 Usability and Its Connection to the

Black Box Problem

The questionnaire results show that a significant

percentage of respondents view AI tools as” user-

friendly, quick," and significantly better than

traditional development methods, since the mean

response for question 5 was 5. For question 6,

Leveraging Artiﬁcial Intelligence Tools in Digital Media Development to Lower Development Barriers

561

however, which asked about the statement “I

understand how the interaction functions work,” the

mean responses were significantly lower at 2.5. This

difference highlights the widespread “black box

problem” in AI tools: people can use sophisticated

models without having an idea of the underlying

principles and limitations that drive them.

One promising pathway to addressing the black-

box problem is the adoption of Explainable Artificial

Intelligence (XAI). Mathew emphasizes that XAI is

evolving toward enhancing model interpretability and

user cognitive understanding, with the primary aim of

making model decision-making processes more

transparent (Mathew, 2025). Bauer further finds that

providing feature-based explanations significantly

improves users’ comprehension and information-

building capabilities, thereby altering their trust and

usage attitudes toward AI systems. These insights

suggest that integrating XAI elements into AI-driven

media tools, such as ml5.js and Ready Player Me,

could not only improve usability but also strengthen

users’ cognitive control and trust in the creative

process (Bauer, K., 2023).

The imbalance between the "enthusiasm for

application" and the "lack of awareness" can lead to

greater adoption of the technology by creators and, in

the process, compromise their capacity to apply the

required changes for reasons of either complex

demands or restrictions from the tools. It is important

to recognize this point in both educational and

professional environments.

5.3 the Implications for Digital Media

Education

Results from this experiment show two major

implications for pedagogical procedures:

First is artificial intelligence tools could provide

an accelerated verification system for teaching

methods and learning experiences.

People who have minimal experience or

backgrounds other than engineering can make use of

tools involving artificial intelligence to expedite the

construction of interactive prototypes over a short

period, shortening the amount of time normally

needed, indeed, a factor of significant value for

innovation and discovery work.

Second is the continuation of the dimension of

"technical understanding" is also critical.

Over-reliance upon the use of tools reduces the

ability of experienced users to cope with complex

technical nuances. Therefore, it becomes necessary

for machine learning tools to include unambiguous

explanations of algorithmic concepts in pedagogical

environments so that learners can grasp essential

recognition processes and data-flow reasoning and

thus develop into effective "system designers" and

not just as "tool operators."

5.4 Considerations for Future Tool

Development and Industry Practice

Relative to industry procedures, the fading barriers to

the use of tools for artificial intelligence will

increasingly promote the democratization of design

procedures, consequently making a larger audience

for nontechnical creatives to work on projects focused

on digital media and interaction design. The

development is expected to promote a revolution in

team dynamics, where artists and designers handle

tasks concerning prototyping, and tech creators focus

on refining complex systems.

However, the situation requires additional

requirements in the creation of artificial intelligence

tools:

Improving transparency: Increased visual

feedback during model runs is crucial for supporting

users in understanding the underlying functional

principles of the tool.

Modularization involves providing "plug-and-

play" modules suited for beginners, as well as for

others who already have a grounding, to tune

parameters with increased specificity. Educational

congruity: Convey pedagogical examples and first-

hand activities focused on developing innovative

learning to minimize the learning process duration.

5.5 Constraints and Potential

Directions for Future Research

The constraints on the results are the comparatively

small sample size, comprising only two nonspecialist

participants, and the relatively simple tasks used. As

a future direction, possible expansion can be made in

a number of different ways:

Recruiting more diverse user groups (e.g.,

different professional backgrounds or age groups);

Including increasingly complex task scenarios,

which involve multimodal interaction and interaction

in mixed reality environments;

Integration of vast observations is needed to study

the learning trajectories and competencies

development of the users while working on tools

embedded with artificial intelligence.

Artificial intelligence technology has greatly

reduced the obstacles to the initial creation of digital

media interactive technologies, but this feeling of

"empowerment" is limited, since it mostly focuses on

EMITI 2025 - International Conference on Engineering Management, Information Technology and Intelligence

562

rapid implementation and prototyping. While AI is a

solid helper for creators, it by no means eliminates the

need for technical reasoning and creative control.

Educators and tool developers need to strike a balance

between "user-friendliness" and "transparency."

6 CONCLUSION AND

PROSPECTS FOR FURTHER

RESEARCH

6.1 Research Summary

This study investigated whether non-experts can

effectively use AI tools, specifically ml5.js and Ready

Player Me, to lower development barriers in digital

media interaction. Two participants without

programming backgrounds were asked to complete

gesture recognition and virtual character control

tasks, with data collected through observation,

surveys, and semi-structured interviews. The results

show that AI tools significantly reduced the difficulty

of implementing these functions, allowing

participants to complete tasks—typically requiring

advanced technical skills—quickly and without

substantial external support. However, despite their

operational success, participants demonstrated

limited understanding of the underlying system

principles and tended to treat the tools as "black

boxes." This indicates that AI tools reduce the

operational threshold but not the cognitive threshold,

and thus do not fully replace the need for technical

comprehension.

6.2 Responding to the Research

Questions

The findings respond directly to the three research

questions. First, non-experts could complete gesture

recognition and virtual character tasks independently

or with minimal assistance, confirming that AI tools

can support productive outcomes without deep

technical skills. Second, while most functions were

accessible through graphical interfaces and API calls,

basic programming knowledge was still required to

understand parameter structures and logical

workflows. Third, participants rated the tools highly

in usability but low in comprehension of underlying

model logic, suggesting a gap between functional use

and conceptual understanding.

6.3 Significance and Practical

Application of Research

This study provides empirical evidence to address the

scholarly question of how AI empowers nontechnical

creators. For education, AI tools can serve as

effective entry points for beginners in interaction

design, provided they are accompanied by instruction

in basic technical reasoning to maintain users'

creative control. The results highlight tool developers'

need to design platforms with more precise feedback

mechanisms and greater model transparency to

enhance users' cognitive control. The findings

suggest that future workflows may increasingly adopt

a "creator × tool platform" model in the creative

industries. AI tools function as collaborative partners,

reshaping production processes and expanding

participation.

6.4 Research Limitations and Future

Directions

Several limitations must be acknowledged. The small

sample size limits the generalizability of the findings,

and the tasks designed for this study were

intentionally simplified, excluding more complex

scenarios involving multimodal interaction,

collaborative networks, or persistent data handling.

Furthermore, the study was short-term and did not

assess participants' skill development over time.

Future research should address these gaps by

expanding the sample size and diversity, introducing

tasks of greater complexity, and conducting

longitudinal studies to explore the long-term effects

of AI tool use on technical competence, learning

motivation, and creativity. Additionally, further work

could investigate the integration of speech

recognition, natural language processing, and other

multimodal AI technologies to extend the scope of

AI-assisted digital media interaction.

7 CONCLUSIONS

The advent of technological innovations fueled by

artificial intelligence is drastically transforming the

ground of digital media creation from a developer-

centric model to a creator-centric model. This has

significantly reduced the technical barriers that once

restricted new creators, enabling them to use

sophisticated interactive media technologies without

requiring in-depth programming knowledge. The

findings of this study suggest that AI-augmented tools

Leveraging Artiﬁcial Intelligence Tools in Digital Media Development to Lower Development Barriers

563

not only enhance creativity horizons but also enhance

efficiency at all stages of development, thus enabling

users to focus more on design thinking and creating

innovative content.

Despite these advantages, a considerable gap

exists between usability and a thorough cognitive

understanding of the underlying technologies. This

means that while artificial intelligence tools have the

potential to supplement the skills of authors, they

need to go hand-in-hand with pedagogical efforts

focused on enhancing technical and critical literacy.

In short, this research highlights the importance of

working with AI-powered tools as an insightful

turning point and not a definitive result. Introducing

such tools to real-world applications and teaching

methodologies might develop a more diverse and

creative environment for creativity and foster further

experimentation, interdisciplinarity, and the

continuous evolution of interactive digital media.

Future research activity could investigate the

long-term effects of outputs from artificial

intelligence on the development of user

competencies, independence, and novel

methodologies. Comparison of studies across

different cultural and disciplinary backgrounds could

produce a more sophisticated understanding of these

tools' reception, adoption, and reinterpretation across

broad user populations.

REFERENCES

Bauer, K., von Zahn, M., & Hinz, O. (2023). Expl (AI) ned:

The impact of explainable artificial intelligence on

users’ information processing. Information Systems

Research, 34(4), 1582-1602.

Billinghurst, M., Clark, A., & Lee, G. (2015). A survey of

augmented reality. Foundations and Trends® in

Human–Computer Interaction, 8(2-3), 73-272.

Burrell, J. (2016). How the machine ‘thinks’:

Understanding opacity in machine learning algorithms.

Big Data & Society, 3(1), 2053951715622512.

Lee, H., & Ko, Y. (2022). Integrating AI tools in design

education: Opportunities and challenges. International

Journal of Art & Design Education, 41(1), 122–135.

Lugaresi, C., Tang, J., Nash, H., McClanahan, C., Uboweja,

E., Hays, M., ... & Grundmann, M. (2019). Mediapipe:

A framework for building perception pipelines. arXiv

preprint arXiv:1906.08172.

Manovich, L. (2020). AI and the future of creativity. AI &

Society, 35(2), 173–182. Retrieved from

https://api.pageplace.de/preview/DT0400.9781000924

787_A46597486/preview-

9781000924787_A46597486.pdf

Mathew, D. E., Ebem, D. U., Ikegwu, A. C., Ukeoma, P. E.,

& Dibiaezue, N. F. (2025). Recent emerging techniques

in explainable artificial intelligence to enhance the

interpretable and understanding of AI models for

human. Neural Processing Letters, 57(1), 16.

McCarthy, L. (2019). ml5.js: Friendly Machine Learning

for the Web. Retrieved from https://ml5js.org

[Accessed 7 Aug. 2025].

Upadhyaya, N. (2023). Low-Code/No-Code platforms and

their impact on traditional software development: A

literature review. No-Code Platforms and Their Impact

on Traditional Software Development: A Literature

Review (March 21, 2023).

Wachs, J. P., Kölsch, M., Stern, H., & Edan, Y. (2011).

Vision-based hand-gesture applications.

Communications of the ACM, 54(2), 60-71.

Wolf3D. (2021). Ready Player Me: Cross-game Avatar

Platform. Retrieved from https://readyplayer.me

[Accessed 7 Aug. 2025].

Yan, Z. (2021). The impacts of low/no-code development

on digital transformation and software development.

arXiv preprint arXiv:2112.14073.

EMITI 2025 - International Conference on Engineering Management, Information Technology and Intelligence

564