Використання генеративного штучного інтелекту (ГШІ) в наукових дослідженнях
The emergence and growing capabilities of Generative Artificial Intelligence (GAI) are profoundly transforming scientific research. Although AI extends human intelligence by automating certain tasks, it complements rather than replaces human creativity. This article discusses the implications of AI...
Gespeichert in:
| Datum: | 2024 |
|---|---|
| 1. Verfasser: | |
| Format: | Artikel |
| Sprache: | Englisch |
| Veröffentlicht: |
The National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute"
2024
|
| Schlagworte: | |
| Online Zugang: | https://journal.iasa.kpi.ua/article/view/315269 |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| Назва журналу: | System research and information technologies |
| Завантажити файл: |  |
Institution
System research and information technologies| _version_ | 1866302983824736256 |
|---|---|
| author | Petrenko, Anatolii |
| author_facet | Petrenko, Anatolii |
| author_sort | Petrenko, Anatolii |
| baseUrl_str | http://journal.iasa.kpi.ua/oai |
| collection | OJS |
| datestamp_date | 2024-11-16T18:06:34Z |
| description | The emergence and growing capabilities of Generative Artificial Intelligence (GAI) are profoundly transforming scientific research. Although AI extends human intelligence by automating certain tasks, it complements rather than replaces human creativity. This article discusses the implications of AI for the scientific process, including ethical considerations and the need for a balanced approach that combines the strengths of human and artificial intelligence in the process of discovering knowledge and solving complex problems. The discussion extends to the need for universities to adapt their curricula to prepare future researchers for the AI era, emphasizing scenario-based thinking and uncertainty management as important skills for the future. |
| doi_str_mv | 10.20535/SRIT.2308-8893.2024.3.08 |
| first_indexed | 2025-07-17T10:28:37Z |
| format | Article |
| fulltext |
Publisher IASA at the Igor Sikorsky Kyiv Polytechnic Institute, 2024
Системні дослідження та інформаційні технології, 2024, № 3 133
TIДC
МЕТОДИ, МОДЕЛІ ТА ТЕХНОЛОГІЇ ШТУЧНОГО
ІНТЕЛЕКТУ В СИСТЕМНОМУ АНАЛІЗІ
ТА УПРАВЛІННІ
UDC 004.89:004.3
DOI: 10.20535/SRIT.2308-8893.2024.3.08
THE ROLE OF GENERATIVE ARTIFICIAL INTELLIGENCE
(GAI) IN SCIENTIFIC RESEARCH
A.I. PETRENKO
Abstract. The emergence and growing capabilities of Generative Artificial Intelli-
gence (GAI) are profoundly transforming scientific research. Although AI extends
human intelligence by automating certain tasks, it complements rather than replaces
human creativity. This article discusses the implications of AI for the scientific
process, including ethical considerations and the need for a balanced approach that
combines the strengths of human and artificial intelligence in the process of discov-
ering knowledge and solving complex problems. The discussion extends to the need
for universities to adapt their curricula to prepare future researchers for the AI era,
emphasizing scenario-based thinking and uncertainty management as important
skills for the future.
Keywords: Generative Artificial Intelligence, hypothesis analysis and testing,
searching scientific data sources, research planning, writing and editing scientific
manuscripts, organizing and presenting results.
THE FUNCTIONS OF GSI SERVICES IN SCIENTIFIC RESEARCH
INVESTIGATION
At the end of last year 2023 a number of leading Ukrainian universities (including
Igor Sikorsky Kyiv Polytechnic Institute, Kherson State University, Zaporizhzhia
Polytechnic National University and others) developed and adopted recommenda-
tions on the use of artificial intelligence in learning, teaching and research [1–3].
Based on these recommendations, we will try to detail and illustrate the possibili-
ties and consequences of using Generative Artificial Intelligence (GAI) in con-
ducting research and presenting the results, taking into account the aspects of
Open Science [4].
Based on the experience gained, it is recommended to use artificial intelli-
gence in research for the following activities:
automated generation of hypotheses and concepts based on data analysis;
analysis and testing of hypotheses and iteration of research processes;
finding sources of scientific data, reviewing, interpreting and citing them;
searching for and extracting specific data from large databases, which
significantly speeds up the information retrieval process;
building a plan/structure for a project/qualifying thesis;
A.I. Petrenko
ISSN 1681–6048 System Research & Information Technologies, 2024, № 3 134
gathering and processing information related to the topic of the pro-
ject/assignment;
analyzing and processing large data sets, identifying patterns, templates
and correlations in them;
translating, editing and summarizing texts;
organizing and comparing the results obtained;
bridging the gaps between different fields of research by providing expla-
nations and combining concepts from different disciplines by finding common
trends and patterns;
writing scientific reports;
writing articles (grammar, translation, paraphrasing, summarizing) ac-
cording to the requirements of scientific journals;
automatically generate graphs, charts and other visual representations of
data, including video, to illustrate key findings and trends;
plagiarism checking;
converting spoken information into printed text (natural language processing);
summary generation.
These activities can be considered as functions of the respective GAI ser-
vices. They require changes in the training of future scientists and professionals
and are a powerful means of personalizing education by adapting content and ex-
periences in ways that were previously impossible. Developing future-oriented
skills in future scientists, such as scenario thinking, systems thinking, and manag-
ing uncertainty and complexity, requires more than memorizing or even managing
large data sets.
THE GENERATION AND TESTING OF RESEARCH HYPOTHESES FORM
THE BASIS OF SCIENTIFIC RESEARCH
GAI is already influencing the development of science, going beyond simple
automation and becoming an active participant in the pursuit of knowledge and
understanding. However, in scientific research, the use of artificial intelligence
represents a significant shift in paradigm, enabling active collaboration between
machines and humans to formulate research hypotheses and questions. Artificial
intelligence systems have traditionally been powerful tools for data analysis.
However, their evolution now enables them to generate hypotheses based on pat-
terns that may escape human observation. GAI algorithms can sift through mas-
sive amounts of data much quicker than humans, identifying interesting patterns
or unique connections. This can lead to new hypotheses for scientists to investi-
gate further. The combination of machine-driven hypotheses and human experi-
ence shows promise for scientific advancement. It is important to maintain human
control over the scientific process and interpretation of results.
However, the ability of artificial intelligence to generate hypotheses raises
questions about the nature of creativity in research. Can artificial intelligence tru-
ly exhibit creativity in hypothesizing, or are they restricted to pattern recognition
in available data?
Additionally, the intersection of artificial intelligence and research extends
beyond hypothesis generation to encompass the formulation of research ques-
tions. Active collaboration between machines and researchers can expand the
The role of Generative Artificial Intelligence (GAI) in scientific research
Системні дослідження та інформаційні технології, 2024, № 3 135
scope and depth of scientific research by identifying gaps in knowledge and sug-
gesting profound questions. This approach emphasizes the indispensable role of
human understanding in forming the research program. Although GAI has great
potential, integrating it into hypothesis formation is not without issues. One sig-
nificant problem is the “black box” nature of many advanced GAI algorithms.
The “black box” issue arises as the complex nature of GAI makes it difficult to
understand how it generates ideas. AI can be compared to a chef who creates a
secret recipe without disclosing the ingredients. To trust AI, it is important to un-
derstand its thought process, which can be challenging.
As these systems become more complex, it becomes increasingly difficult
for human researchers to understand the reasoning behind the non-traditional hy-
potheses generated by the GAI. Large language models such as Copilot (ChatGPT-4)
(https://www.bing.com/chat), Gemini Ulta (https://gemini.google.com/) or Cloud-3
(https://claude.ai/) can exacerbate this issue. Figure 1 contains a fragment of Gem-
ini’s answer on the question about examples of GAI-powered hypotheses in Com-
puter Science.
The results of talk on the same topic with ChatGPT-4 and Cloud-3 are
shown on Fig. 2 and Fig. 3 respectively.
Generative AI (GAI) has gained prominence in various fields, including computer
science. Here are a few hypotheses related to GAI applications in this domain:
Enhanced Code Generation: GAI models, such as Generative Pretrained Trans-
formers (GPT), could be leveraged to automatically generate high-quality code snippets,
improving software development efficiency and reducing human effort.
Data Augmentation: GAI techniques can create synthetic data samples, augmenting
existing datasets for training machine learning models. This could enhance model
generalization and performance.
Natural Language Understanding: GAI models excel at understanding and
generating natural language. They could be used to improve chatbots, virtual assistants,
and sentiment analysis systems.
Drug Discovery: GAI can assist in drug discovery by generating molecular
structures with desired properties, potentially accelerating the search for new medications.
Art and Creativity: GAI-generated art, music, and literature could inspire new
creative works or even lead to novel artistic styles.
Fig. 2. Illustration of talk with ChatGPT-4 about its ability to generate hypotheses
Fig. 1. Illustration of talk with Gemini about its ability to generate hypotheses
Gemini Advanced
Hypotheses on GAI EnhancingSpecific Tasks:
1. Automated Code Generation: GAIs will outperform the best human pro-
grammers in generating error-free, efficient code across multiple programming
languages when given a clear problem specification.
2. Network, Security: GAI-powered intrusion detection systems will have a sig-
nificantly lower false-positive rate and higher detection rate of novel cyberat-
tracks compared to current state-of-the-art systems.
3. Natural Language Understanding: GAI models will achieve human-level
performance on complex natural language understanding tasks like reading
comprehension, text summarization, and nuanced sentiment analysis.
4. Data Compression: GAI-designed data compression algorithms will consis-
tently achieve higher compression ratios with minimal loss of information co-
pared to existing methods on diverse data types.
A.I. Petrenko
ISSN 1681–6048 System Research & Information Technologies, 2024, № 3 136
Here are some concrete hypotheses related to potential applications of Artificial
General Intelligence (AGI) in computer science:
AGI systems could autonomously develop new algorithms and data structures
that are more efficient and scalable than those designed by human programmers.
AGI systems could seamlessly integrate and optimize different areas of com-
puter science, such as machine learning, computer vision, natural language processing,
and robotics, leading to more sophisticated and versatile applications.
AGI systems could automatically debug and optimize legacy software systems,
improving their performance, security, and maintainability.
AGI systems could dynamically allocate and manage computing resources (e.g.,
CPU, memory, storage) across distributed systems, leading to more efficient utilization of
hardware resources.
AGI systems could autonomously explore and generate novel computational
architectures, programming paradigms, and programming languages that are better
suited for specific tasks or domains.
AGI systems could automatically translate high-level specifications or require-
ments into efficient and robust software implementations, facilitating rapid prototyp-
ing and development of complex systems.
AGI systems could continuously monitor and adapt software systems to chang-
ing environments, user behaviors, and security threats, enabling self-healing and self-
optimizing systems.
AGI systems could autonomously discover and exploit patterns and insights in
large-scale data, leading to breakthroughs in areas such as bioinformatics, computational
finance, and scientific computing.
AGI systems could dynamically synthesize and integrate knowledge from mul-
tiple domains, enabling the development of interdisciplinary applications that leverage
insights from diverse fields.
AGI systems could autonomously explore the vast search space of potential
software solutions, leading to the discovery of novel and innovative approaches to com-
plex computational problems.
These hypotheses span various areas of computer science, including algorithm de-
sign, software engineering, distributed systems, programming languages, and interdisci-
plinary applications. However, it’s important to note that the realization of these hypothe-
ses would depend on significant advancements in AGI research and the development of
robust and reliable AGI systems.
Fig. 3. Illustration of talk with Cloud-3 about its ability to generate hypotheses
which is the most meaningful
It seems that GAI can’t actually generate a new thought, rather than repeat
something from the training array, or substitute the most likely answer from the
same training array. There is a very thin line between creating new thoughts (hy-
potheses) and just looking for information from an existing array. The lack of in-
terpretation and concretization can hinder the acceptance of hypotheses generated
by artificial intelligence in the scientific community. In addition, the datasets on
which GAI models are trained can contain inherent prejudices that may influence the
generated hypotheses. If this bias is not addressed, it can distort the development of
perspectives in science. It is important to recognize that although artificial intelligence
can process vast amounts of information, it cannot discern subtleties [5; 6].
Nevertheless, at the beginning of October 2023, researchers gathered in
Stockholm, who under led by Hiroaki Kitano, a biologist and CEO of Sony AI,
The role of Generative Artificial Intelligence (GAI) in scientific research
Системні дослідження та інформаційні технології, 2024, № 3 137
considered questions about the introduction of Nobel prizes for developers of arti-
ficial intelligence and artificial cooperation intelligence and people who made an
outstanding scientific contribution [7]. And it’s fair, because GAI is on its own
influence on the development of scientific research can be rightly considered as a
quality fundamental direction of modern science.
As GAI plays a more active role in hypothesis formation, ethical considera-
tions become a priority. The responsible use of GAI requires constant vigilance to
prevent undesirable consequences. Researchers must be vigilant in detecting and
mitigating biases and one-sidedness in the training datasets, ensuring that the sys-
tem of artificial intelligence does not preserve or strengthen existing inequalities
in the adequacy of generated solutions for various branches and tasks.
In addition, the ethical implications of hypotheses generated by GAI, par-
ticularly in sensitive areas such as genetics or social sciences, require careful con-
sideration. Transparency in the decision-making process regarding GAI hypothe-
ses is crucial for building trust within the scientific community and society as a
whole. Striking the right balance between innovation and ethical responsibility is
an ongoing challenge that demands constant attention, as collaboration between
humans and GAI continues to evolve [8]. Accurate thinking, creativity, and un-
derstanding of context play a vital role in improving and testing hypotheses gen-
erated by the GAI.
However, despite their intelligence, GAI cannot replace human scientists. It
is still necessary for people to carefully consider the suggestions made by AI and
interpret their true meaning. Good teamwork among different experts, such as
computer scientists, ethicists, and researchers, is essential to ensure the best use of
AI. Researchers must act as critical evaluators, questioning the assumptions made
by artificial intelligence algorithms and ensuring that the proposed hypotheses are
consistent with available knowledge. This will not only improve our understand-
ing of the underlying processes but also ensure that the hypotheses meet ethical
and scientific standards.
The integration of artificial intelligence into the hypothesis generation proc-
ess is an ongoing journey with enormous potential. The combined efforts of hu-
mans and machines hold the promise of accelerating scientific discovery, generat-
ing new ideas and solving complex problems facing humanity. However, this
journey requires a balanced approach that recognizes the strengths of artificial
intelligence while respecting the unique skills
and ethical considerations that humans bring to
the table.
For example, the transformative power of
GAI in hypothesis generation is changing the
landscape of scientific research. However, this
would not be possible without a joint and
dynamic partnership between humans and ma-
chines, which has the potential for unprece-
dented advances and opens up a new era of
scientific research and understanding that is
inherent in Industry 5.0 technologies (Fig. 4).
Fig. 4. Symbol of human-machine
partnership generated by GAI
A.I. Petrenko
ISSN 1681–6048 System Research & Information Technologies, 2024, № 3 138
PRIORITY SERVICES FOR STUDENTS TO LEARN MODERN RESEARCH
METHODOLOGY
With the increasing use of GAI in scientific research, it is important for universi-
ties to prepare future researchers to work in this era. While some may argue that
the use of artificial intelligence in fundamental scientific research is too complex
for higher education purposes, it is still important for students to learn modern
research methodology. It may be argued that not all students are capable of under-
standing modern advanced research methods. Empirical research can be con-
ducted through student projects to test their scientific development abilities and
the consequences of using artificial intelligence in the learning process. Modern
university curriculums worldwide include some traditional aspects of scientific
research.
The use of GAI in scientific research involves various aspects, including ex-
perimentation, data collection and interpretation, modelling, and automation of
design. Therefore, the question of prioritizing certain GAI services for inclusion
in educational programs arises. Developing educational adaptations of profes-
sional GAI services can help teach future scientists about the ethics of using GAI
in scientific research and instil in them a sense of responsibility. However, teach-
ers should not only be trained to use artificial intelligence tools and data, but also
to understand the rapidly changing landscape of modern science in the era of arti-
ficial intelligence. It is important to develop individual solutions for each spe-
cialty to ensure the appropriate selection of basic GAI services.
Regulatory documents of universities [1–3] provide recommendations for
using existing GAI services, which now number in the thousands. It is important
to note that there are many options available for choosing GAI services. GAI ser-
vices have become popular due to their ability to democratize GIS (Geographic
information systems) applications, making them accessible to people without
technical training. Each GAI service learns from a specific set of data that deter-
mines its effectiveness in performing specialized tasks. It is crucial to accurately
determine and select the service that best meets your needs to get the most out of
these GAI services. As a first step in implementing university policies for the use
of artificial intelligence in academic activities, we have selected and demonstrated
five effective interdisciplinary GAI services below. Their selection was based on
an attempt to ensure their uniqueness and to avoid describing many of the GAI
services that can be found in most publications. They serve as tools to support
different aspects of the research process.
Semantic scholars. Search and Discovery of Scientific Information
Semantic Scholar (https://www.semanticscholar.org/) is a GAI service that pro-
vides access to a huge database of more than 211 million articles from all fields of
science, making it one of the largest repositories of scientific literature. This com-
prehensive collection covers a wide range of subjects including: physics, chemis-
try, biology, medicine, history, computer science, etc. From cutting-edge research
papers to historical publications, Semantic Scholar has a wealth of resources to
The role of Generative Artificial Intelligence (GAI) in scientific research
Системні дослідження та інформаційні технології, 2024, № 3 139
support researchers in their pursuit of knowledge. One of the key features of Se-
mantic Scholar is its powerful search and filtering capabilities based on artificial
intelligence.
The service uses natural language processing and machine learning algo-
rithms to analyze the content of articles and extract relevant information. This
allows users to perform advanced searches by specific keywords, authors, jour-
nals, publication dates, and other parameters, enabling accurate and efficient arti-
cle retrieval [10]. The selection of literature sources on the topic of this article
made by the service is illustrated in Fig. 5.
It is interesting to compare received results with those obtained from a
Google search engine based on the user’s query. Often, the data obtained from
such searches do not correspond to the search goal and return simple URLs of
sites and fragments of content that are not always relevant. Developers must
check website content, filter out irrelevant information, and optimize it according
to constraints. Fig. 5 shows that the Semantic Scholar’s results differ from
Google’s. It mainly consists of published articles, each providing detailed infor-
mation on its significance (through the number of citations) and direction (topic).
The article can be sorted by different criteria, including the number of authors and
sources used to build it (References), related papers such as images and publica-
tions (Related Papers), publication date (Recency), number of citations (Citation
Count), importance (Most Influential Papers), and relevance to the topic (Rele-
vance).
Fig. 5. A fragment of selected literature (1940 sources in English and 3 in Ukrainian)
A.I. Petrenko
ISSN 1681–6048 System Research & Information Technologies, 2024, № 3 140
Please note that this service has 56 alternative purpose-specific services,
such as Scholarcy, OpenRead, Elicit, Scispace, Scite, Research Buddy, Mirror-
think, and Epsilon. A comparison of their results with Semantic Scholar is avail-
able on the website https://theresanaiforthat. com/ai/semanticscholar/.
Explain a paper. Extracting and understanding information from sources
Explain Paper (https://www.explainpaper.com/) is an AI service designed to
work with scientific publications from any field, including natural sciences, social
sciences, humanities, etc. However, it works best with materials that are complex
in technical and scientific jargon. The explanations aim to cover the main con-
cepts and key findings as accurately as possible. However, it is always advisable
to refer to the original article for technical details. You should think of an AI ex-
planation as a “tutor friend” pointing out the main ideas. The Explain Paper ser-
vice accepts PDF documents from any source — journals, preprint servers, uni-
versity websites, etc. You can set the depth of explanation from basic to very
detailed (expert). This service is designed for researchers and scientists seeking to
broaden their knowledge beyond their respective fields.
The Explain GAI service is designed for:
researchers and scientists who want to extend their knowledge beyond
their own field. Artificial intelligence explanations help them to quickly under-
stand key ideas in papers from other fields;
graduate students and researchers who need to process large volumes of
dense, complex documents, as simplified explanations allow for faster reading;
industry professionals in technical positions who want access to the latest
research and ideas, as the service makes the most recent article more accessible;
lifelong learners and knowledge seekers who want to keep up to date, be-
cause Explain Paper makes complex documents accessible.
Downloading the content of Gemini’s answer about examples of AI-powered
hypotheses in Computer Science (Fig. 2) and selecting part of the text to get the
GAI-generated explanation is shown in Fig. 6.
According to the ExplainPaper website, users can save time and improve the
accuracy of their work by using this service [11]. This service has 63 alternatives,
including Ask Your PDF, Brevity, HeyScience, Skimlt.ai, Summatity, Summary-
Box, Docu-Ask, SciSpace Copilot, AI Next Assistant, Doks.ai, PDFAI.io, and
others (https://topai.tools/alternatives/explainpaper).
The role of Generative Artificial Intelligence (GAI) in scientific research
Системні дослідження та інформаційні технології, 2024, № 3 141
F
ig
. 6
. H
ig
hl
ig
ht
a
s
ec
tio
n
of
te
xt
in
a
n
ar
tic
le
a
nd
it
s
ex
pl
an
at
io
n
A.I. Petrenko
ISSN 1681–6048 System Research & Information Technologies, 2024, № 3 142
IntelliConsult. Project consulting/engineering
IntelliConsult (https://intelliconsultai.com/) uses the power of artificial intelli-
gence to analyze project data and provide valuable information and recommenda-
tions for project optimization [12]. It can be used to easily create project plans,
allocate resources and set realistic timeframes. IntelliConsult is easy to use thanks
to its user-friendly interface. It guides you through the entire consulting process,
making it easy to access the features and support you need. Fig. 7 shows the input
interface for entering a description of the project for which the plan was devel-
oped (Fig. 8).
Fig. 7. The input interface for entering the project description
Fig. 8. Illustration of the results of developing individual strategies and justifying project
decisions
The role of Generative Artificial Intelligence (GAI) in scientific research
Системні дослідження та інформаційні технології, 2024, № 3 143
The IntelliConsult service is a guide for achieving excellence in projects. It
offers strategic project management and planning, risk analysis and management,
project result optimization, and inspiration for innovation. This service provides
project planning and strategic support throughout the entire project life cycle,
thanks to the latest artificial intelligence technology.
According to https://theresanaiforthat.com/ai/intelliconsult /, IntelliConsult
has 43 competitors, including Wolfe, Gitsul Group, AI-Engageme, Consultant in,
Strategic Mind, ProfitGPT, My Consultant, Peter Drucker, ai revolution,
ConsultGPT, and AI Transform.
Academic GPT. Writing scientific papers
AcademicGPT (https://academicgpt.net/ ) is an GAI service that helps profession-
als with the complex process of writing scientific papers [13]. Its artificial intelli-
gence algorithms excel at generating annotations and concise abstracts, providing
users with a significant increase in productivity. This service allows professionals
to upload their papers in PDF format and use the power of GAI algorithms to im-
prove their work. However, it is important to remember that while AcademicGPT is a
powerful service, it should complement, not replace, thoughtful research writing.
First, a user uploads a draft of a research paper to AcademicGPT and selects
one of three sections: Write, Feedback or Explain. In the drop-down menu
“Choose which type of paper section the AI should write” (Fig. 9), you can select
the type of task for which AcademicGPT should create content.
Next, in the “Feedback” and “Explain” sections, the user selects clear
instructions or key points that AcademicGPT should follow when creating the
content of his or her paper. In the Feedback section, the following instructions can
be selected: Strength of Argument, Language, Coherence, Originality, Citation,
and in the Explain section: Simple and Short, Simple and Long, Complex and
Long, Complex and Short.
After entering the instructions, clicking on the Create button allows
AcademicGPT to use its GAI capabilities to generate an abstract or any other
appropriate academic section based on the requirements. Fig. 10 shows an
example of an abstract for this article “The use of generative artificial intelligence
(GAI) in scientific research”.
Fig. 9. Selecting a task in the “Write” section: abstract, conclusions, summary, future
manuscript sections, ethical considerations section, online presentation, LaTeX format)
A.I. Petrenko
ISSN 1681–6048 System Research & Information Technologies, 2024, № 3 144
If necessary, users can review or expand the generated content by reusing the
AcademicGPT service and providing additional instructions or adjusting the initial
input. Once the user is satisfied with the generated content, they can export it in
the required format or copy and paste it into their research paper.
It is important to note that there are many alternative services available for
this purpose, such as Shutterstock, Grammarly, Gamma AI, CrushOn.AI, Run-
wayml, You, DeepAI, PixAI - AI Art Generator, MaxAI.me, and Simplified.
A comparison of their results with AcademicGPT is provided on the website
https://www.toolify.ai/alternative/academicgpt.
SlidesGPT. Presentation/Publication of results
SlidesGPT (https://slidesgpt.com/) is a GAI service that makes it easy to create
slides for Google presentations, for example, the results of a research paper. To
create a slide, you need to prepare a presentation plan or content source, then
simply paste the text and get a presentation in seconds. It is possible to review the
draft slides and make corrections or additions. Instead of just starting with a
prompt, you can also ask the service to convert a PDF or web page into a presen-
tation outline. With an intuitive interface, the service understands the user’s pres-
entation needs and helps them design slides that are both informative and attrac-
tive. It is ideal for professionals who need to quickly create powerful
presentations, teachers who want to make educational content more engaging, or
students preparing for academic projects.
Fig. 10. Illustration of the abstract been generated by GAI
The role of Generative Artificial Intelligence (GAI) in scientific research
Системні дослідження та інформаційні технології, 2024, № 3 145
Fig. 11 shows 4 out of 10 slides generated by SlidesGPT using the annota-
tion from this article in Ukrainian.
After viewing the generated presentation, the user has the opportunity to cus-
tomize it according to their specific needs. They can edit the content, replace im-
ages, rearrange slides and make any necessary changes. SlidesGPT’s GAI service
is capable of creating doctoral-level presentations, including detailed, factual and
specific information. It can integrate user-provided data into presentations,
ensuring that the source data is accurate and up-to-date. Among the many al-
ternative presentation services, MagicSlides[14], SlidesAI, AhaSlides, SlidesGo,
Beautiful AI, Invideo, Canva, Tome, Hitch, Gamma, Prezi and Syis
(https://ahaslides.com/blog/slides-ai-platforms/) are worth mentioning.
These five GAI services are valuable for both students and teachers, as they
can enhance productivity and work quality while reducing the time and effort re-
quired for important tasks. In the fast-changing technological landscape and with
the growing demands on the competencies of future professionals, clarity is espe-
cially crucial.
CONCLUSIONS
Generative Artificial Intelligence (GAI) is transforming scientific research. GAI
offers a range of services to optimize hypothesis generation and concept devel-
Fig. 11. Illustration of generated slides on the topic “The use of Generative Artificial
Intelligence (GAI) in scientific research”
A.I. Petrenko
ISSN 1681–6048 System Research & Information Technologies, 2024, № 3 146
opment, facilitate hypothesis analysis and testing, assist in finding and extracting
scientific data, assist in research design planning, manage large data sets, perform
translations and editing, meet journal publication requirements, create data visu-
alizations, check for plagiarism, transcribe audio recordings, and generate sum-
maries [15]. Expanding research capacity in this way improves personalized edu-
cation by shifting from rote learning to critical thinking and problem-solving
[8; 9]. Future research in this field should aim to develop methodologies for
evaluating GAI-generated hypotheses that are both innovative and scientifically
rigorous. One potential avenue for exploration is the design of GAI systems with
explainable algorithms, which would offer transparency in their reasoning proc-
esses and make their hypotheses more understandable and acceptable to the scien-
tific community.
The use of Generative Artificial Intelligence (GAI) in research and academia
has both significant benefits and challenges. Tools such as AcademicGPT and
SlidesGPT demonstrate the potential to empower people by simplifying the crea-
tion of scientific documents and presentations. With the help of AI services such
as Semantic Scholar and Explain Paper, researchers can increase their access to
knowledge, improve understanding and develop new hypotheses. However, de-
spite these benefits, the implementation of GAI should be approached with cau-
tion, taking into account ethical considerations and the risk of data bias. Scientific
institutions need to develop and refine guidelines for the responsible use of GAI,
ensuring that it serves as a complement to human intelligence, not a replacement
[16–19]. Many universities mentioned in [18] are equipped by GAI detection
tools (e.g. GPTZero, Turnitin, GPTKit, Winston AI, etc.) [20]. By striking a bal-
ance between innovation and maintaining academic integrity, GAI can become a
powerful ally in the pursuit of scientific progress and the evolution of educational
practices.
REFERANCES
1. Policy on the use of artificial intelligence for academic activities at Igor Sikorsky
KPI. Available: https://osvita.kpi.ua/sites/default/files/downloads/polityka-
vykorystannia-shtuchnogo-intelektu_2023.pdf
2. The policies for using artificial intelligence in education, teaching, and research
at Kherson State University. Available: https://www.kspu.edu/ File-
Download.ashx?id=00653012-555c-46b2-bb64-05ba9bf26773
3. The principles for using generative artificial intelligence at Zaporizhzhia Poly-
technic National University. Available: https://zp.edu.ua/uploads/pubdocs/2023/
Dekl_pro_pryn_vykor_heneratyvnoho_shtuchnoho_intelektu.pdf
4. M. Marienko, V. Kovalenko, “Artificial Intelligence and Open Science,” Physical
and Mathematical Education, vol. 38, no. 1, pp. 48–53, 2023. doi: 10.31110/2413-
1571-2023-038-1-007
5. “AI is transforming how science is done. Science education must reflect this change,”
SCIENCE, vol. 382, issue 6677, 21 Dec 2023. doi: 10.1126/science.adm9788
6. Uttkarsha Bhosale, AI-Driven Hypotheses: Real world examples exploring the poten-
tial and challenges of AI-generated hypotheses in science. Available:
https://www.enago.com/academy/ai-generated-research-hypothesis/
7. Tim Keary, Artificial Superintelligence (ASI). Available: https://www. techope-
dia.com/definition/31619/artificial-superintelligence-asi
8. Generative AI in Academic Research: Perspectives and Cultural Norms. Available:
https://research-and-innovation.cornell.edu/generative-ai-in-academic-research/
The role of Generative Artificial Intelligence (GAI) in scientific research
Системні дослідження та інформаційні технології, 2024, № 3 147
9. Yueqiao Jin, Lixiang Yan, Vanessa Echeverria, Dragan Gašević, and Roberto Marti-
nez-Maldonado, Generative AI in Higher Education: A Global Perspective of Insti-
tutional Adoption Policies and Guidelines, arXiv:2405.11800v1 [cs.CY], 20 May
2024. Available: https://arxiv.org/pdf/2405.11800
10. Semantic Scholar. Available: https://www.wikidata.uk-ua.nina.az/Semantic_Scholar.html
11. Enhance Your Research Paper Understanding with Explain Paper. Available:
https://www.toolify.ai/ai-news/enhance-your-research-paper-understanding-with-
explain-paper-1069228
12. IntelliConsult. Available: https://www.stork.ai/ru/ai-tools/intelliconsult
13. TopAI.tools. Available: https://topai.tools/t/academicgpt
14. Magic Slides. Available: https://www.popularaitools.ai/tools/magic-slides
15. Best Practices for Generative AI in Research. Available: https://www.aje.com/arc/
best-practices-generative-ai-in-research/
16. F. Miao, W. Holmes, Guidance for generative AI in education and research. 2023.
Accessed on: February 5, 2024. Available: https://unesdoc.unesco.org/
ark:/48223/pf0000386693
17. L. Yan et al., “Practical and ethical challenges of large language models in educa-
tion: A systematic scoping review,” British Journal of Educational Technology, 55,
pp. 90–112, 2024.
18. B.L. Moorhouse, M.A. Yeo, and Y. Wan, “Generative AI tools and assessment:
Guidelines of the world’s top-ranking universities,” Computers and Education
Open, 5, 100151, 2023. Available: https://www.sciencedirect.com/science/ arti-
cle/pii/S2666557323000290
19. M. Sullivan, A. Kelly, and P. McLaughlan, “ChatGPT in higher education: consid-
erations for academic integrity and student learning,” J. Appl. Learn. Teach., 6 (1),
2023. Available: https://doi.org/10.37074/jalt.2023.6.1.17
20. 9 Best Free AI Detectors for Teachers in 2024. Available: https://www.classpoint.
io/blog/best-free-ai-detectors-for-teachers
Received 30.05.2024
INFORMATION ON THE ARTICLE
Anatolii I. Petrenko, ORCID: 0000-0001-6712-7792, Educational and Research Institute
for Applied System Analysis of the National Technical University of Ukraine “Igor
Sikorsky Kyiv Polytechnic Institute”, Ukraine, e-mail: tolja.petrenko@gmail.com
ВИКОРИСТАННЯ ГЕНЕРАТИВНОГО ШТУЧНОГО ІНТЕЛЕКТУ (ГШІ)
В НАУКОВИХ ДОСЛІДЖЕННЯХ / А.І. Петренко
Анотація. Поява та зростаючі можливості генеративного штучного інтелекту
(ГШІ) глибоко трансформують наукові дослідження. Хоча ГШІ розширює
людський інтелект шляхом автоматизації певних завдань, він радше доповнює,
ніж замінює людську креативність. Розглянуто наслідки ГШІ для наукового
процесу, включаючи етичні міркування та необхідність збалансованого підхо-
ду, який об’єднує сильні сторони людського та штучного інтелекту в процесі
відкриття знань і вирішення складних проблем. Обговорення поширюється на
необхідність для університетів скорегувати навчальні програми для підготовки
майбутніх дослідників до епохи ГШІ, підкреслюючи сценарне мислення та
управління невизначеністю як важливі навики майбутнього.
Ключові слова: генеративний штучний інтелект, аналіз і тестування гіпотез,
пошук джерел наукових даних, планування дослідження, написання і редагу-
вання наукових рукописів, упорядкування і презентація результатів.
|
| id | journaliasakpiua-article-315269 |
| institution | System research and information technologies |
| keywords_txt_mv | keywords |
| language | English |
| last_indexed | 2025-07-17T10:28:37Z |
| publishDate | 2024 |
| publisher | The National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute" |
| record_format | ojs |
| resource_txt_mv | journaliasakpiua/aa/6ee5f44be7bf5739fe39b92ac11ab3aa.pdf |
| spelling | journaliasakpiua-article-3152692024-11-16T18:06:34Z The role of generative artificial intelligence (GAI) in scientific research Використання генеративного штучного інтелекту (ГШІ) в наукових дослідженнях Petrenko, Anatolii генеративний штучний інтелект аналіз і тестування гіпотез пошук джерел наукових даних планування дослідження написання і редагування наукових рукописів упорядкування і презентація результатів Generative Artificial Intelligence hypothesis analysis and testing searching scientific data sources research planning writing and editing scientific manuscripts organizing and presenting results The emergence and growing capabilities of Generative Artificial Intelligence (GAI) are profoundly transforming scientific research. Although AI extends human intelligence by automating certain tasks, it complements rather than replaces human creativity. This article discusses the implications of AI for the scientific process, including ethical considerations and the need for a balanced approach that combines the strengths of human and artificial intelligence in the process of discovering knowledge and solving complex problems. The discussion extends to the need for universities to adapt their curricula to prepare future researchers for the AI era, emphasizing scenario-based thinking and uncertainty management as important skills for the future. Поява та зростаючі можливості генеративного штучного інтелекту (ГШІ) глибоко трансформують наукові дослідження. Хоча ГШІ розширює людський інтелект шляхом автоматизації певних завдань, він радше доповнює, ніж замінює людську креативність. Розглянуто наслідки ГШІ для наукового процесу, включаючи етичні міркування та необхідність збалансованого підходу, який об’єднує сильні сторони людського та штучного інтелекту в процесі відкриття знань і вирішення складних проблем. Обговорення поширюється на необхідність для університетів скорегувати навчальні програми для підготовки майбутніх дослідників до епохи ГШІ, підкреслюючи сценарне мислення та управління невизначеністю як важливі навики майбутнього. The National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute" 2024-09-28 Article Article Peer-reviewed Article application/pdf https://journal.iasa.kpi.ua/article/view/315269 10.20535/SRIT.2308-8893.2024.3.08 System research and information technologies; No. 3 (2024); 133-147 Системные исследования и информационные технологии; № 3 (2024); 133-147 Системні дослідження та інформаційні технології; № 3 (2024); 133-147 2308-8893 1681-6048 en https://journal.iasa.kpi.ua/article/view/315269/306082 |
| spellingShingle | генеративний штучний інтелект аналіз і тестування гіпотез пошук джерел наукових даних планування дослідження написання і редагування наукових рукописів упорядкування і презентація результатів Petrenko, Anatolii Використання генеративного штучного інтелекту (ГШІ) в наукових дослідженнях |
| title | Використання генеративного штучного інтелекту (ГШІ) в наукових дослідженнях |
| title_alt | The role of generative artificial intelligence (GAI) in scientific research |
| title_full | Використання генеративного штучного інтелекту (ГШІ) в наукових дослідженнях |
| title_fullStr | Використання генеративного штучного інтелекту (ГШІ) в наукових дослідженнях |
| title_full_unstemmed | Використання генеративного штучного інтелекту (ГШІ) в наукових дослідженнях |
| title_short | Використання генеративного штучного інтелекту (ГШІ) в наукових дослідженнях |
| title_sort | використання генеративного штучного інтелекту (гші) в наукових дослідженнях |
| topic | генеративний штучний інтелект аналіз і тестування гіпотез пошук джерел наукових даних планування дослідження написання і редагування наукових рукописів упорядкування і презентація результатів |
| topic_facet | генеративний штучний інтелект аналіз і тестування гіпотез пошук джерел наукових даних планування дослідження написання і редагування наукових рукописів упорядкування і презентація результатів Generative Artificial Intelligence hypothesis analysis and testing searching scientific data sources research planning writing and editing scientific manuscripts organizing and presenting results |
| url | https://journal.iasa.kpi.ua/article/view/315269 |
| work_keys_str_mv | AT petrenkoanatolii theroleofgenerativeartificialintelligencegaiinscientificresearch AT petrenkoanatolii vikoristannâgenerativnogoštučnogoíntelektugšívnaukovihdoslídžennâh AT petrenkoanatolii roleofgenerativeartificialintelligencegaiinscientificresearch |