Julio-Diciembre 2015

Vol. 5 No. 2

Octubre-diciembre 2024

Vol. 14 No. 3

Interacción y Perspectiva. Revista de Trabajo Social Vol. 14 No3 / octubre-diciembre, 2024

620

Interacción y Perspectiva Dep. Legal pp 201002Z43506

Revista de Trabajo Social ISSN 2244-808X

Octubre-diciembre

ARTÍCULO DE INVESTIGACIÓN

Aprendizaje basado en las preferencias humanas: un estudio piloto sobre la

percepción de los estudiantes sobre el uso de la IA y ChatGPT

DOI: https://doi.org/10.5281/zenodo.11154746

Dalina Dumitrescu *

Resumen

El estudio examinó la conciencia, comprensión y actitudes de los estudiantes hacia la

inteligencia artificial (IA) y su impacto en la gestión del proceso educativo. Se utilizó una

encuesta con preguntas abiertas. El objetivo era analizar la percepción de los

encuestados sobre la importancia, uso e impacto de la IA, incluyendo ChatGPT, en el

ámbito laboral y educativo en el que participan. Se aplicó a 25 cursantes de maestría en

la Facultad de Finanzas y Banca de la Universidad de Estudios Económicos de Bucarest.

Las conclusiones se basaron en el procesamiento de las respuestas abiertas utilizando

una herramienta de IA. Los hallazgos revelaron que todos los encuestados eran

conscientes de la presencia de la IA. Las principales fuentes de información sobre la IA

fueron el lugar de trabajo, las redes sociales y los amigos. El análisis del sentimiento de

las respuestas mostró un coeficiente positivo más alto con respecto al impacto de la IA

en la educación. Aunque se reconoce que este estudio piloto tiene limitaciones, las

principales conclusiones indican que los estudiantes están receptivos a la sociedad

digitalizada. El uso de plataformas como ChatGPT y software de IA en actividades

prácticas es limitado. Valoran con cautela la experiencia de trabajar con generadores de

texto de IA. El predominio de sentimientos positivos relacionados con el uso de ChatGPT

tanto en el lugar de trabajo como en las actividades educativas proporciona una base

sólida para la integración formal, coherente, eficiente y participativa de la IA en enfoques

y procesos educativos responsables.

Palabras clave: educación superior, IA; impacto de la IA; ChatGPT; Digitalización y

gestión educativa.

Abstract

Learning based on human preferences: A pilot study regarding the student's

perception of the AI and the use of ChatGPT

The study examined the awareness, understanding, and attitudes of students towards

artificial intelligence (AI) and its impact on educational management. A survey with

open-ended questions was used for data collection. The objective was to analyze the

respondents' perception of the importance, use, and impact of AI, including ChatGPT, in

their workplace and educational settings. It was administered to 25 master's students

Dalina Dumitrescu / Aprendizaje basado en las preferencias humanas: un estudio piloto sobre la percepción

de los estudiantes sobre el uso de la IA y ChatGPT

621

at the Faculty of Finance and Banking, University of Economic Studies in Bucharest. The

conclusions were based on processing the open-ended responses using an AI tool. The

findings revealed that all respondents were aware of the presence of AI. The main

sources of information about AI were the workplace, social media, and friends. The

sentiment analysis of the responses showed a higher positive coefficient regarding the

impact of AI on education. While acknowledging the limitations of this pilot study, the

main conclusions indicate that students are receptive to the digitalized society. The use

of platforms like ChatGPT and AI software in practical activities is limited. They cautiously

value the experience of working with AI text generators. The predominance of positive

sentiments related to the use of ChatGPT in both the workplace and educational activities

provides a solid foundation for the formal, coherent, efficient, and participatory

integration of AI into responsible educational approaches and processes.

Keywords: higher education, AI; impact of AI; ChatGPT; Digitization and education

management.

Recibido: 18/03/2024 Aceptado: 29/04/2024

* Universidad de Estudios Económicos de Bucarest, Rumania, ORCID ID: https://orcid.org/0000-0002-4633-

1791 E-mail: dalina.dumitrescu@fin.ase.ro

1. Introduction

In the last 10 years, artificial intelligence (AI) has developed a potentially disruptive

impact on all human society. Artificial Intelligence (AI) refers to reproducing human

ability to get and use knowledge and skills by machines, particularly computer systems.

These include learning (acquiring knowledge and skills through experience), reasoning

(concluding available information), problem-solving (finding solutions to complex

challenges), and decision-making (selecting the best course of action based on analysis).

AI aims to create systems that mimic human cognitive abilities, enabling them to

perform activities that typically require human intelligence. Here are some key areas

where AI has made and continues to make a significant impact:

TABLE Nro. 1

KEY AREAS WHERE AIHAS A SIGNIFICANT

Application Area

References

Key Points

Industries

Revolutionization via

AI Automation

Jarret et al.

(2021), Sarmah &

Shekhar (2019)

Streamlining processes, reducing human

intervention, and enhancing efficiency

(Automation and Efficiency)

AI in Medicine

Saraswat et al.

(2022), Yu et al.

(2018)

Assisting in medical diagnosis, drug

discovery, personalized treatment plans,

and predictive analytics (Healthcare)

AI in Finance

Milana et al.

(2021), Shamima

et al. (2022)

Analyzing financial data, informing trading

decisions, detecting fraud, and predicting

market trends (Finance)

Interacción y Perspectiva. Revista de Trabajo Social Vol. 14 No3 / octubre-diciembre, 2024

622

AI in Manufacturing

Nti et al. (2022),

Yao et al. (2017)

Optimizing production processes,

enhancing quality control, and enabling

predictive maintenance of machinery

(Manufacturing)

AI in Transportation

Prashant et al.

(2022), Bharadiya

(2023)

Revolutionizing transportation, reducing

accidents and congestion, and improving

efficiency (Transportation)

AI in Entertainment

Fu et al. (2002),

Nader et al.

(2022)

Content recommendation, computer-

generated graphics, animations, and

enhanced video game experiences

(Entertainment)

AI in Natural

Language Processing

Chowdhary

(2020), Raina et

al. (2022)

Enabling chatbots, virtual assistants,

accurate language translation, and

revolutionizing interactions with

technology (Natural Language Processing)

AI in Global

Connectivity

Leyva-Mayorga et

al. (2022),

Akhavan et al.

(2020)

Facilitating communication across

language barriers, fostering global

collaboration and understanding (Global

Connectivity)

AI in Environmental

Sustainability

Tsolakis et al.

(2022), Feroz et

al. (2021)

Climate modeling, analyzing environmental

data, optimizing energy consumption, and

managing natural resources efficiently

(Environmental Sustainability)

AI in Personalization

in Marketing

Chandra et al.

(2022), Gao et al.

(2022)

Analyzing consumer behavior, offering

personalized product recommendations,

and executing targeted marketing

campaigns (Personalization in Marketing)

AI in Cybersecurity

Ansari et al.

(2023), Zhang et

al. (2022)

Detecting cyber threats, identifying

anomalies, and strengthening security

measures against data breaches and

attacks (Cybersecurity)

AI in Agriculture

Liu (2020),

Sharma (2021)

Precision farming, optimizing irrigation,

monitoring crop health, and predicting

yield outcomes (Agriculture)

AI in Social Impact

Shangyao et al

(2007), Linardos

et al. (2022),

Lee et al. (2023),

Chiou et al. (2022)

Significant contributions in disaster

response, analyzing large-scale data

during emergencies, and assisting in

resource allocation for relief efforts (Social

Impact)

AI in Education

Alam (2022)

Liu et al. (2023)

Sung-Hee et al.

(2023)

Transforming education through

personalized learning experiences,

intelligent tutoring systems, automated

grading, and adaptive learning platforms

(Education)

Source: constructed by the author in dialog with ChatBPT (2023)

In essence, AI's significance lies in its ability to process and analyze massive volumes

of data, make predictions, and perform tasks that were previously considered beyond

Dalina Dumitrescu / Aprendizaje basado en las preferencias humanas: un estudio piloto sobre la percepción

de los estudiantes sobre el uso de la IA y ChatGPT

623

the capabilities of machines. It is reshaping industries, enhancing human capabilities,

and driving innovation in countless ways.

While the transformative impact of AI has been largely positive, there are also

concerns about ethical considerations, job displacement, privacy issues, and biases in AI

algorithms. Balancing these challenges while harnessing AI's potential for positive

change is a critical ongoing endeavor.

2. Specific background information

In the last 15 years, the main stakeholders and quality accreditation agencies for

higher education had a specific objective to assess the introduction of technology in the

education process, in general, and at the level of individual courses in special Bakir

(2016), Hodges et al. (2022)

Incorporating technology into the classroom has been a positive step forward for

education. Using laptops and dedicated sites to store documents, searching the web for

resources, and utilizing statistical software such as Excel, Stata, R, and Python have all

been beneficial in the social sciences. Additionally, simulations and games can help

transfer knowledge in an engaging way. As we move towards in-person learning again,

it's important to consider how we can continue to integrate technology effectively and

thoughtfully while preserving the value of face-to-face interaction. The adoption and the

extent of implementation of education technology were highly related to the resources

of the university (approved budget to buy the technological infrastructure in the social

fields where the teaching process could have been realized in a classical way without

technology). The ability and openness of teachers to identify, design, and adopt new

elements of educational technology in the classroom is a great challenge, especially

where teachers' area of expertise is not modern technology.

The historical development of AI in education reflects a journey from early theoretical

concepts to practical applications that are transforming how learning is delivered,

monitored, and personalized. As AI technologies become more sophisticated, the

potential for enhancing educational experiences and outcomes continues to expand.

Table 2. Summarizes the key developments in AI's role in education across different

decades.:

TABLE Nro. 2

AI'S ROLE IN EDUCATION ACROSS DIFFERENT DECADES

Time Period

Key Developments

Examples and Applications

1950s-1960s

Emergence of AI as a field of

study, Turing Test proposed by

Alan Turing

Early programs: "Logic Theorist,"

"Geometry Theorem Prover"

1970s-1980s

Development of knowledge-

based systems and expert

systems, MYCIN aids in

diagnosing infectious diseases

Cognitive science and AI research

influence intelligent tutoring systems

(ITS), e.g., Soar, ACT-R

Interacción y Perspectiva. Revista de Trabajo Social Vol. 14 No3 / octubre-diciembre, 2024

624

1990s

Rise of Intelligent Tutoring

Systems (ITS), examples like the

Geometry Tutor and AutoTutor

World Wide Web emergence facilitates

web-based learning, AI used for adaptive

e-learning environments

2000s

Advances in machine learning

and data analytics enhance ITS

capabilities, educational games

and simulations with AI

integration

Games like "Zoombinis" and "Math

Blaster" incorporate AI for engaging

learning experiences

2010s

Rise of Massive Open Online

Courses (MOOCs) with AI scaling

instruction, automated

assessments, and personalized

recommendations

Introduction of chatbots and virtual

assistants like IBM's "Jill Watson" for

student assistance

2020s and

Beyond

Continued role of AI in education,

supporting adaptive learning,

learning analytics, and real-time

feedback systems

Natural language processing and

sentiment analysis gauge student

emotions and engagement levels,

addressing ethical concerns

2022

November 30, 2022 Chat GPT

was launched

ChatGPT employs deep learning, a

branch of machine learning, to generate

text that closely resembles human

language using transformer neural

networks. The transformer model

anticipates text outcomes, such as the

following word, sentence, or paragraph,

by leveraging patterns observed in its

training data.

The training process initiates with

general data and progressively

transitions to more specialized data

aligned with a particular task. ChatGPT

initially undergoes training with diverse

online text to grasp the nuances of

human language. Subsequently, it

refines its conversational skills by

learning from transcripts, honing its

ability to engage in meaningful

dialogues.

Source: Constructed based on literature review by the author in dialog with

ChatBPT (2023)

The limitations and the new rules imposed by the COVID crisis forced the adoption

of technology in the education process to a new level of adoption and implementation.

Given the forced and sudden lockdown of the entire social life, to maintain the

educational process uninterrupted during the pandemic, the universities looked to the

experience of the total online programs as a rescue solution to accomplish their mission.

Dalina Dumitrescu / Aprendizaje basado en las preferencias humanas: un estudio piloto sobre la percepción

de los estudiantes sobre el uso de la IA y ChatGPT

625

The aim was to ensure for the graduates the exposure and experience regarding the

whole curriculum, the core professional set of knowledge, skills, and attitude as close as

possible to normal time. In other words, to attain learning goals and objectives as in the

face-to-face educational process. Starting with 2020 a lot of debates were in the main

flow of literature related to the pros and cons related to in-person and virtual learning,

As a result, Universities struggled to build strong virtual learning communities as new

online education models are going beyond the replication of face-to-face activities, see

Webinar: Unlocking the Power of AI: How Tools Like ChatGPT Can Make Teaching Easier

and More Effective-HBP, May 2, 2023.

However, the COVID pandemic did not mean only the adoption of online education

models with already calibrated technology that coexisted previously with face-to-face

ones. Technology has advanced and opened generously to the public with tools, models,

and products, especially in the area of artificial intelligence (AI).

The tools and products offered on a large scale and often free of charge by AI, and

recently developed Artificial Intelligence Text Generators (AITG), with application for

research work and implicitly also in educational activity (reports, essays, finalization

works), they represent completely new facilities with profound implications, Bahroun

(2023).

Using AI, the online research of information and documentation sources is much

simplified the part of conception, structuring and elaboration of research works (essays,

final papers doctoral thesis) including the innovation and personalization components is

done automatically. Finding automatic answers to the questions to quizzes or field exams

including science exams is more and more accurate. Thus, the human factor's

contribution and work based on his/her inherent knowledge, skills, and abilities is no

longer relevant.

The tasks to be realized by the students such as admission essays, research papers,

field essays and reports, innovative code lines writing, and final papers (such as

undergraduate, dissertation, and Ph.D. thesis) supposed to the personal contributions

based on the accumulated knowledge, skills, and attitudes. The individual contributions

and the personal innovative approach were built by educational experiences at courses

and trans-curricular levels and were the base for assessing the learning and grading,

Holland (2023).

The code of ethics in all the universities states strong rules for the ethics in using

online resources and plagiarism (similitude check).

In this aim, in many universities introduced for checking the effective enforcement

of integrity rules in the academic intellectual contributions and assessing the knowledge,

different systems (such as shields in place to limit the access of the students to the web

pages for the in-campus exams) or antiplagiarism softs- similitude check (for the final

papers).

In the current period, an extremely important question is to what extent AI should

be allowed, adopted, or limited/reduced/prohibited in the process of building and

Interacción y Perspectiva. Revista de Trabajo Social Vol. 14 No3 / octubre-diciembre, 2024

626

consolidating the fund of knowledge and skills for young people, in the process of forming

the professional bases of independent work.

ChatGPT was released in 2019 by OpenAI, followed by an updated version of this

conversational chatbot in late November 2022 based on Reinforcement Learning with

Human Feedback (RLHF).

ChatGPT is algorithm-driven and aims to generate human-like text that can be used

in conversation. As a functionality, it can answer questions, provide information, and

participate in natural language discussions with users. ChatGPT is used as a customer

service chatbot, virtual assistant, and conversational interface for websites or mobile

applications. It can also be used to generate content for social media or create chatbot

scripts for marketing or entertainment purposes. It is a powerful tool for creating

chatbots that can have intelligent and engaging conversations with users, - see Webinar-

Will AI Replace the Educator- HBP August 10, 2023,

As for educational applications, it is mentioned by (Hwang et al., 2020) that there

are several roles of AI in education, such as serving as an intelligent tutor, guidance,

learning tool/policy partner, or advisor.

Chen et al. (2020) found that research in the field of Artificial Intelligence in

Education AIEd rapidly developed around 2012, and the number of articles, grants, and

citations in the field is still increasing. The significant increase in citations received by

AIEd studies demonstrates the wider impact and influence of AIEd research in academia.

The findings are consistent with those from the main flow of research results (e.g., Chen

et al., 2022, Chu et al., 2022, Chiu et al., 2023)

Recent research studies have significantly contributed to the understanding of how

AIEd and technology can be effectively integrated into educational settings, fostering

personalized learning experiences, improving learning outcomes, and addressing

important ethical and practical considerations.

As we transition back to face-to-face learning, it's important for us to consider how

we can incorporate the benefits of virtual learning communities into our in-person

education. We need to ask ourselves what worked well in the online format and how we

can integrate those elements into the traditional classroom experience. At the same

time, we should also reflect on what we value about face-to-face learning and make sure

we preserve those elements. It's a balancing act, but by taking a thoughtful and

intentional approach, we can create a truly dynamic and effective learning environment.

Important works in educational data mining and learning analytics contributed to

research on using data to understand and improve learning processes (Ryan et al.,

2016).

Another important research line is intelligent tutoring systems, affective computing,

and human-computer interaction in education, (Woolf et al., 2009), the development of

cognitive tutors and intelligent tutoring systems mainly applies cognitive science

principles to education (Aleven et al., 2006).

Dalina Dumitrescu / Aprendizaje basado en las preferencias humanas: un estudio piloto sobre la percepción

de los estudiantes sobre el uso de la IA y ChatGPT

627

3. Description of the gap in our knowledge that the study was designed to fill

Mijwil et al. (2020), studied the impact of artificial intelligence applications, their

implication in learning, and how they can be employed in the future development of the

teaching/learning process. They concluded that artificial intelligence is a key factor of

growth in future digital societies and must be properly exploited to build a new world

where future graduates must be competitive.

Ömer Osmanoǧlu et al. (2020), collected student feedback from a university was

utilized to evaluate course materials. The machines were employed to categorize the

materials based on positive, negative, or neutral sentiments regarding each teaching

resource. Subsequently, the course materials were enhanced by addressing the negative

feedback.

Nikolić et al. (2020), started from student course assessment conducted at sentence-

level analysis to identify one or more aspects in the sentences and classify their polarity

into positive or negative sentiment.

Misuraca et al. (2021), Gkontzis et al. (2020) concluded that in an educational

context, Opinion Mining allows processing students’ comments and creating powerful

analytics.

Shaik et al. (2023), concluded that through advancements in sentiment techniques

and AI methodologies, student comments can depict their sentiment orientation with

minimal human intervention. The significance of emotional analysis in education was

examined across four levels: document level, sentence level, entity level, and aspect

level. In conjunction with sentiment annotation techniques, the study delved into the

role of AI in sentiment analysis using methodologies such as machine learning, deep

learning, and transformers. The findings highlight the influence of sentiment analysis on

educational processes, aiming to improve pedagogy, decision-making, and evaluation.

As we transition back to face-to-face learning, it's important for us to consider how

we can incorporate the benefits of virtual learning communities into our in-person

education. We need to ask ourselves what worked well in the online format and how we

can integrate those elements into the traditional classroom experience. At the same

time, we should also reflect on what we value about face-to-face learning and make sure

we preserve those elements. It's a balancing act, but by taking a thoughtful and

intentional approach, we can create a truly dynamic and effective learning environment.

How do we educate using responsible the new technology mainly the one generated by

AI and AITG specifically? The question is going beyond the simple one what the

implications of the use of laptops and smartphones during class will be? It is about

accepting or resisting using ChatGPT to do assignments and examinations. If accepted,

what to do as responsible educators is not to diminish the quality of the learning process

and attain properly the objectives assumed by the missions of our institutions.

In a recent webinar held by HBP were presented the results of a survey from June

2023 applied in multiple countries (UK, US, Germany, Spain, Australia, Nederland,

Interacción y Perspectiva. Revista de Trabajo Social Vol. 14 No3 / octubre-diciembre, 2024

628

France) with the question “How is your school using the AI?” The 12 schools that

responded formulated the answers at 3 levels of responsibility and implication:

discussion, at the faculty level, and the school level. The main conclusion was that most

educational institutions are at the early stages of using AI in their activity. Currently,

they are in a period of exploration. The initiatives are mainly tactical and incremental,

and the main uses of AI are material creation and AI chatbots, see Webinar-

Experimenting with AI in the Classroom-Harvard School of Education- September 6,

2023

4. The study objective

The work, a pilot study, investigates to what extent the students in the master

programs are aware of artificial intelligence, from which sources they got information

about AI, and specifically about ChatGPT. We also investigated their opinion about the

impact of AI on jobs and education, if they used ChatGPT, and for what purpose. Those

questions are also the research questions of the study and the procession of the answers

to them are presented in detail. According to our knowledge, there is no study

investigating the perception and the sentiments of students regarding the AI influence

at work and in education.

The positive effects of investigating students’ opinions and sentiments are supported

by research studies. For example, Chakraborty et al. (2020) used as an instrument for

research a survey in which undergraduate students in an Indian university expressed

their opinions on online education during the pandemic. The conclusions were that the

students learn better in physical classrooms (65.9%) than through online education, and

they felt that the teachers improved their online teaching. The students gave favorable

evaluations to the software and online study materials employed as support. However,

they expressed concerns about the stress associated with online education, noting its

impact on their health and social life.

The paper is structured as follows: The methodology part will describe the approach

and the tools used for the analysis of the answers, the next two parts will present the

results and the discussion of the result. The conclusion will close the study.

5. Methodology

As presented in the introduction, most educational institutions are currently in the

early stage of incorporating AI into their instructional activities.

From this perspective, the interest arose to investigate, in the form of a pilot study,

whether at the level of the faculty where I work, the master's program students -

predominantly employees - are familiar with the concept of AI. The study aims to explore

from which sources they gather information about AI, and how they perceive the impact

of AI on their professional activities, as well as on educational endeavors. Specifically,

regarding ChatGPT, the students were questioned about their usage of it, their

Dalina Dumitrescu / Aprendizaje basado en las preferencias humanas: un estudio piloto sobre la percepción

de los estudiantes sobre el uso de la IA y ChatGPT

629

experience with working hours involving it, and the specific activities for which they

employed it. To the best of our knowledge, there is currently no such study conducted

that delves into students' opinions regarding the impact of Artificial Intelligence Text

Generators (AITG) on education, research, and the workplace.

At the end of May 2023, a questionnaire was distributed to the master's program

students at the Faculty of Finance and Banking. The questionnaire was administered in

paper format to the 25 students who were in attendance. Responses to the

questionnaire, in English, were anonymous, and the collection process was conducted

randomly. The purpose of gathering this information was to communicate to the

students, and the response rate was 100%.

The questionnaire comprises 5 questions, out of which 4 are open-ended questions

and one is quantitative. Two out of the 5 questions of the survey were inspired by the

questions addresses online to the educated participants during the Webinar: Unlocking

the Power of AI: How Tools Like ChatGPT Can Make Teaching Easier and More Effective-

HBP, May 2, 2023.

The preparation of the survey is based on literature review to ensure that the

questions are grounded in existing research and theories as well as the logic in the

investigation of the awareness and use of new concepts/products services by individuals,

In (2017).

Given the relatively short time after the release of the CHATGDP – only 6 months-

the current study was developed as a pilot study, with a small sample of respondents to

to identify potential issues with question wording, clarity, ambiguity, or issues with the

questions addressed. Conducting the pilot study was intended to save time and to obtain

valuable insights in an area where no other information was available. Some validation

conditions were observed: (a) Ensure that questions are clear, concise, and easily

understood by the target audience; (b) Ensure that questions are neutral and do not

lead respondents to a particular answer; (c) Avoid wording that might introduce bias or

influence respondents; (d) Gather feedback from respondents after the questionnaire

has been administered, Mellinger et al. (2020).

The research questions of the study are aligned with the questions of the survey as

follows: If the students are aware of artificial intelligence, and from which sources do

they get information about AI (Q1); What is their opinion about the impact of AI on

current jobs (Q2); What is their opinion about the impact of AI on education, (Q3); If

they used ChatGPT- as working hours (Q4); For what purpose they used ChatGPT (Q5)

Considering the investigated area of interest, the decision was made to utilize open-

ended questions.

The option for open-ended questions was determined by the arguments that the free-

form gives space to respondents to answer in an open-text format and the freedom to

express himself/herself in as much (or as little) detail as they prefer. Open-ended

questions help to see the respondent’s perspective, as the feedback is in their own words

instead of preformatted answers. The free format allows getting more meaningful

Interacción y Perspectiva. Revista de Trabajo Social Vol. 14 No3 / octubre-diciembre, 2024

630

answers, accurate responses, and a better way to analyze the sentiments of

respondents, Siedlecki, (2022).

The processing of responses for the open-ended questions was carried out using the

AI Monkey Learn tool, while the quantitative question was processed by the author using

Excel.

AI text analysis employs natural language processing (NLP) to automate the

classification and extraction of data from texts. This approach is highly effective,

particularly in automatically analyzing open-ended survey responses. From the site

Monkey Learn, it was extracted the relevant characteristics of the program. The machine

learning models it uses can apply several pre-trained text analysis models that can

process the survey responses very rapidly on the next dimensions (from the MonleyLearn

site):

“Sentiment Analyzer: automatically analyzes text for opinion polarity

(positive, negative, neutral).

NPS Feedback Analyzer: automatically sorts NPS responses by Ease of Use,

Features, Pricing, and Support.

Keyword Extractor: extracts the most used and most important words from your

survey responses.

Company Extractor: automatically extracts the names of businesses and

organizations from surveys or any text.”- Excerpt from Monkey Learn site.

https://monkeylearn.com/blog/survey-analysis/

Given the purpose of the pilot study and the reduced complexity of the data collected,

the survey answers were analyzed with a Sentiment Analyzer and keyword extractor.

For each open question processed with Monkey Learn, the paper presents the overall

picture of the processing of the answers as well as the graphic representation of the

focused subsections. For the quantitative question- question 4-, the graph of the results

processed by the author manually is also presented.

6. Results

The main results of the analysis of survey answers are presented in line with the

research question of the paper investigated by the survey questions. For question 1 will

be presented all the elements obtained from the AI processing: The synthesis image

(dashboard), the result from the keywords extractor (aspect count and Keywords cloud),

the result for sentiment analyzer (positive-green, neutral orange, and negative–red) and

the overall sentiment. The results of questions 2, 3 and 5 will be presented as the results

of processing for the two dimensions- keywords extractor and sentiment analyzer. The

answers for question 4, quantitative, manually processed, are presented in Fig. 10.

Dalina Dumitrescu / Aprendizaje basado en las preferencias humanas: un estudio piloto sobre la percepción

de los estudiantes sobre el uso de la IA y ChatGPT

639

The author is fully aware that the number of respondents, the method of their

selection, as well as the complexity of the questionnaire, do not allow the results of the

processing to serve as a sufficient basis, by scientific requirements, for the extrapolation

and generalization of conclusions.

However, in a society undergoing profound and extremely rapid change in this

regard, we believe that the results obtained, with a 100% response rate, can serve as

at least an encouraging starting point for the development and expansion of the study.

In conditions where even in a preliminary and somewhat less scientifically rigorous form,

querying students' opinions about AI and investigating their level of experience with text

generators is lacking, these results can be indicative of students' openness, support, and

experience in working with AI. Academic institutions must prioritize implementing

programs that offer all stakeholders, such as students, professors, and doctoral

researchers, comprehensive education, and training on AI. This will ensure responsible

integration of various AI applications into the educational process and promote

continuous and formal reflection on the ethical use of Artificial Intelligence. On

September 6, 2023, the Harvard School of Education hosted a webinar titled

"Experimenting with AI in the Classroom" to explore this subject further.

The results of the pilot study obtained using AI and non-AI techniques in processing

reveal several aspects that can themselves be points of discussion and further

exploration.

It is important to note that respondents are familiar with and knowledgeable about

AI technology.

Regarding the source of information, the top 5 mentioned sources include the

workplace (exclusively in a positive context), news, friends, and scientific articles. The

university as an educational institution is mentioned less frequently but also in a positive

context (see Fig. 2 and Fig. 3). The results show that all the students were aware of the

AI and the main source to find out about it was at work, from news, from friends and

articles.

Concerning the overall sentiment, 76% of responses related to the context in which

they learned about AI were expressed positively, based on sentiment analysis (see Fig.

5).

The perception of AI's influence in the workplace has focused on keywords like

impact, job, work, society, and workplace. The answers express the fact that the

respondents were aware of the impact of AI at work, for their jobs, and in the workplace.

It's noteworthy that sentiment analysis by category highlighted that these keywords

were mostly formulated in positive, but also in negative or neutral contexts. The keyword

"workplace" was expressed 50% in a positive context and 50% in a neutral context (see

Fig. 7).

Regarding the overall sentiment, 69% of responses related to AI's impact on the

current job were expressed positively, according to sentiment analysis.

Interacción y Perspectiva. Revista de Trabajo Social Vol. 14 No3 / octubre-diciembre, 2024

640

The students' perception of AI's influence on education was synthesized by both

machine learning models and non-AI techniques. It's interesting to note the similarity in

results, regardless of the analysis model used, with keywords such as students,

education, impact, positive impact, and task. The non-AI processing model also

highlighted verbs and adjectives describing the actionable and effective aspects of using

AI, such as "will," "can," "use/used," "finding," "think," "helps," or "positive," "faster,"

"easier," and "correct" (see Fig. 8 and Fig. 9).

76.6% of responses regarding the impact of AI on education were expressed

positively from a sentiment analysis perspective), representing the highest percentage

of positive statements in the questionnaire.

The experience of surveyed students in using ChatGPT, Q4, during working hours,

as depicted in Fig.10, reflects a very limited level of experience, mainly at the exploratory

and testing stages. Therefore, 44% of students did not use it at all, which is equal to the

number of those who used it for intervals of time between one hour and 10 hours, with

no student admitting to using it for more than 40 hours.

In terms of usage, the primary activities were information retrieval at the workplace,

finding answers to questions, curiosity about how it works, and generating summaries

(see Fig.11). It's interesting to note the categories detected from responses expressing

only positive sentiments, including order, example, excel, field, database, and

automation of ChatGPT.

Regarding the post-processing CSV document by MonkeyLearn for question 5, 46%

of responses were positive.

8. Conclusion

Despite significant limitations regarding the database used, the pilot study represents

a groundbreaking investigation into awareness of AI's impact on the workplace and

education, as well as the experience in using ChatGPT and the purposes for which it was

used.

The use of AI in processing the questionnaire aimed to ensure consistency between

the investigated subject and the processing tool.

We appreciate that the predominance of positive sentiments regarding the influence

and usage effects of ChatGPT by master's students, both in the workplace and in

educational activities, provides a solid basis for the formal, coherent, efficient, and

responsible integration of AI into the educational process.

Acknowledgement:

In memory of my mother, Zoe (1931-2023)

Dalina Dumitrescu / Aprendizaje basado en las preferencias humanas: un estudio piloto sobre la percepción

de los estudiantes sobre el uso de la IA y ChatGPT

641

Bibliographic references

 Akhavan, M., Ghiara, H., Mariotti, I., & Sillig, C. (2020). “Logistics global network

connectivity and its determinants. A European City network analysis.” Journal

of Transport Geography, 82, 102624.

https://doi.org/10.1016/j.jtrangeo.2019.102624

 Alam, A. (2022). “Employing Adaptive Learning and Intelligent Tutoring Robots

for Virtual Classrooms and Smart Campuses: Reforming Education in the Age of

Artificial Intelligence”. In: Shaw, R.N., Das, S., Piuri, V., Bianchini, M. (eds)

Advanced Computing and Intelligent Technologies. Lecture Notes in Electrical

Engineering, vol 914. Springer, Singapore. https://doi.org/10.1007/978-981-

19-2980-9_32

 Aleven, V., McLaren, B. M., Roll, I., & Koedinger, K. R. (2006). “Toward Meta-

cognitive Tutoring: A Model of help seeking with a Cognitive Tutor”.

International Journal of Artificial Intelligence in Education, 16(2), 101-

128.

 Ansari, M. F., Dash, B., Sharma, P., & Yathiraju, N. (2023). “The Impact and

Limitations of Artificial Intelligence in Cybersecurity: A Literature Review.”

International Journal of Advanced Research in Computer and

Communication Engineering-SSR.

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=4323317

 Bahroun, Z., Anane, C., Ahmed, V., & Zacca, A. (2023). “Transforming

Education: A Comprehensive Review of Generative Artificial Intelligence in

Educational Settings through Bibliometric and Content

Analysis.” Sustainability, 15, 12983. https://doi.org/10.3390/su151712983

 Bakir, N. (2016). “Technology and Teacher Education: A Brief Glimpse of the

Research and Practice that Have Shaped the Field”. TechTrends, 60, 21-29.

https://doi.org/10.1007/s11528-015-0013-4

 Bharadiya, J. (2023).” Artificial Intelligence in Transportation Systems A Critical

Review.” American Journal of Computing and Engineering, 6(1), 34-45.

https://doi.org/10.47672/ajce.1487

 Chakraborty, P., Mittal, P., Gupta, M. S., Yadav, S., & Arora, A. (2020). “Opinion

of students on online education during the COVID-19 pandemic.” Human

Behavior and Emerging Technologies, 3(1).

https://doi.org/10.1002/hbe2.240

 Chandra, S., Verma, S., Lim, W. M., Kumar, S., & Donthu, N. (2022).

“Personalization in personalized marketing: Trends and ways

forward”. Psychology & Marketing, 39, 1529-

1562. https://doi.org/10.1002/mar.21670

 Chen, X., Xie, H., & Hwang, G. J. (2020). “A multi-perspective study on Artificial

Intelligence in Education: grants, conferences, journals, software tools,

Interacción y Perspectiva. Revista de Trabajo Social Vol. 14 No3 / octubre-diciembre, 2024

642

institutions, and researchers.” Computers and Education: Artificial

Intelligence, 1(3),100005. http://dx.doi.org/10.1016/j.caeai.2020.100005

 Chen, X., Zou, D., Xie, H., Cheng, G., & Liu, C. (2022). “Two Decades of Artificial

Intelligence in Education: Contributors, Collaborations, Research Topics, Challenges,

and Future Directions: Educational Technology & Society, 25(1), 28-47.

https://www.jstor.org/stable/48647028

 Chiou, M., Epsimos, G. T., Nikolaou, G., Pappas, P., Petousakis, G., Mühl, S., &

Stolkin, R. (2022). Robot-Assisted Nuclear Disaster Response: “Report and

Insights from a Field Exercise”. IEEE/RSJ International Conference on

Intelligent Robots and Systems (IROS), 4545-4552.

https://doi.org/10.1109/IROS47612.2022.9981881

 Chiu, T. K. F., Xia, Q., Zhou, X., & Chai, C. S., & Cheng, M. (2023). “Systematic

literature review on opportunities, challenges, and future research

recommendations of artificial intelligence in education”. Computers and

Education: Artificial Intelligence, 4(3), 100118.

https://doi.org/10.1016/j.caeai.2022.100118.

 Chowdhary, K. R. (2020). “Natural Language Processing.” In Fundamentals of

Artificial Intelligence. Springer. New Delhi. https://doi.org/10.1007/978-81-

322-3972-7_19

 Chu, H. C., Hwang, G. H., Tu, Y. F., & Yang, K. H. (2022). “Roles and research

trends of artificial intelligence in higher education: A systematic review of the top

50 most-cited articles”. Australasian Journal of Educational Technology,

38(3), 22-42. https://doi.org/10.14742/ajet.7526

 Feroz, A. K., Zo, H., & Chiravuri, A. (2021). “Digital Transformation and

Environmental Sustainability: A Review and Research

Agenda”. Sustainability, 13, 1530. https://doi.org/10.3390/su13031530

 Fu, D., & Houlette, R. (2002). “Putting AI in entertainment: an AI authoring tool

for simulation and games.” IEEE Intelligent Systems, 17(4), 81-84.

https://doi.org/10.1109/MIS.2002.1024756

 Gao, Y., & Liu, H. (2022). “Artificial intelligence-enabled personalization in

interactive marketing: a customer journey perspective.” Journal of Research

in Interactive Marketing, 1-18. https://doi.org/10.1108/JRIM-01-2022-0023

 Gkontzis, A. F., Kotsiantis, S., Kalles, D., Panagiotakopoulos, C. T., &

Verykios, V. S. (2020).” Polarity, emotions and online activity of students and

tutors as features in predicting grades.” Intell. Decis. Technol., 14(3), 409-

436. https://doi.org/10.3233/idt-190137

 Hodges, C. B., Barbour, M., & Ferdig, R. E. (2022). “A 2025 Vision for Building

Access to K-12 Online and Blended Learning in Pre-service Teacher

Education.” Journal of Technology and Teacher Education, 30(2), 201-216.

Dalina Dumitrescu / Aprendizaje basado en las preferencias humanas: un estudio piloto sobre la percepción

de los estudiantes sobre el uso de la IA y ChatGPT

643

Waynesville, NC USA: Society for Information Technology & Teacher

Education. https://www.learntechlib.org/primary/p/221153/

 Holland, B. J. (2023). “ChatGPT 3.5 and 4: Bits Ramifications on Librarianship,

Academia, Education, Publishing, and the Workplace.” In B. Holland

(ed.), Handbook of Research on Advancements of Contactless

Technology and Service Innovation in Library and Information Science

(pp. 316-340). IGI Global. https://doi.org/10.4018/978-1-6684-7693-2.ch016

 Hwang, G. J., Xie, H., Wah, B. W., & Gašević-Vision, D. (2020). “Challenges,

Roles and Research Issues of Artificial Intelligence in Education-Computers and

Education.” Artificial Intelligence, 1, 100001.

 In J. “Introduction of a pilot study.” Korean J Anesthesiol. 2017

Dec;70(6):601-605. doi: 10.4097/kjae.2017.70.6.601. Epub 2017 Nov 14.

PMID: 29225742; PMCID: PMC5716817.

 Jarret, A., Kwang, K., & Choo, R. (2021). “The impact of automation and artificial

intelligence on digital forensics-Wires Forensic Science.” Wiley

Interdisciplinary Reviews. https://doi.org/10.1002/wfs2.1418

 Lee, M., Mesicek, L., Bae, K., & Ko, H. (2023). “AI advisor platform for disaster

response based on big data.” Concurrency and Computation Practice and

Experience, 35(16), e6215. https://doi.org/10.1002/cpe.6215

 Leyva-Mayorga, I., Soret, B., Matthiesen, B., Röper, M., Wübben, D., Dekorsy,

A., & Popovski, P. (2022). “NGSO Constellation Design for Global Connectivity-

Computer Science” Networking and Internet Architecture, Book chapter

Cornel University. https://doi.org/10.1049/PBTE105E

 Linardos, V., Drakaki, M., Tzionas, P., & Karnavas, Y. L. (2022). “Machine

Learning in Disaster Management: Recent Developments in Methods and

Applications.” Machine Learning and Knowledge Extraction, 4(2), 446-473.

https://doi.org/10.3390/make4020020

 Liu S. Y. (2020). “Artificial Intelligence (AI) in Agriculture.” IT Professional,

22(3), 14-15. https://doi.org/10.1109/MITP.2020.2986121

 Liu, M., Yu, D. “Towards intelligent E-learning systems.” Educ Inf Technol 28,

7845–7876 (2023). https://doi.org/10.1007/s10639-022-11479-6

 Marcello, M. M., Machado, I., Magrelli, V., & Dwivedi, Y. K. (2023). “Artificial

intelligence in innovation research: A systematic review, conceptual framework,

and future research directions.” Technovation, 122, 102623.

https://doi.org/10.1016/j.technovation.2022.102623.

 Mijwil, M. M., Aggarwal, K., Mutar, D. S., Mansour, N., & Singh, R. S. S. (2022).

“The Position of Artificial Intelligence in the Future of Education: An Overview.”

Asian Journal of Applied Sciences, 10(2).

https://doi.org/10.24203/ajas.v10i2.6956

Interacción y Perspectiva. Revista de Trabajo Social Vol. 14 No3 / octubre-diciembre, 2024

644

 Mellinger, C. D., & Hanson, T. A. (2020). “Methodological considerations for

survey research:Validity, reliability, and quantitative analysis.” Linguistica

Antverpiensia, New Series: Themes in Translation Studies, 19, 172–190. DOI:

https://doi.org/10.52034/lanstts.v19i0.549

 Milana C, Ashta A.” Artificial intelligence techniques in finance and financial

markets: A survey of the literature.” Strategic Change. 2021; 30: 189–209

Mellinger, https://doi.org/10.1002/jsc.2403

 Misuraca, M., Scepi, G., & Spano, M. (2021). “Using Opinion Mining as an

educational analytic: An integrated strategy for the analysis of students’

feedback.” Studies in Educational Evaluation, 68, 100979.

https://doi.org/10.1016/j.stueduc.2021.100979

 Nader, K., Toprac, P., Scott, S., & Baker, S. (2022).” Public understanding of

artificial intelligence through entertainment media.” AI & Soc.

https://doi.org/10.1007/s00146-022-01427-w

 Nikolić N., Grljević O., & Kovačević A. (2020). “Aspect-based sentiment analysis

of reviews in the domain of higher education,” Electron. Libr., 38(1), 44-

64. https://doi.org/10.1108/el-06-2019-0140

 Nti, I. K., Adekoya, A. F., Weyori, B. A., & Owusu, N. B. (2022). “Applications of

artificial intelligence in engineering and manufacturing: a systematic

review.” Journal of Intelligent Manufacturing, 33(6), 1581-1601.

https://doi.org/10.1007/s10845-021-01771-6

 Ömer Osmanoǧlu, U., Atak O. N., Çaǧlar K., Kayhan, H., & Can, T. (2020).

“Sentiment analysis for distance education course materials: A machine learning

approach.” Journal of Education Technology and Online Learning, 3(1), 31-

48. https://doi.org/10.31681/jetol.663733

 Prashant, S., Zeinab, E., Yui-yip, L., Borowska-Stefańska, M., Wiśniewski, S., &

Maxim, A. (2022). “Dulebenets, Blockchain and AI technology convergence:

Applications in transportation systems,” Vehicular Communications, 38,

00521, ISSN 2214-2096, https://doi.org/10.1016/j.vehcom.2022.100521

 Raina, V., & Krishnamurthy, S. (2022).” Natural Language Processing. In Building

an Effective Data Science Practice.” Apress, Berkeley, CA.

https://doi.org/10.1007/978-1-4842-7419-4_6

 Ryan, S. B., Martin, T., & Rossi, L. M. (2016). “Educational Data Mining and

Learning” Analytics- 16 November.

https://doi.org/10.1002/9781118956588.ch16-

 Saraswat, D., Bhattacharya, P., Verma, A., Prasad, V. K., Tanwar, S., & Sharma,

G. (2022). “Explainable AI for Healthcare 5.0: Opportunities and Challenges.”

IEEE Access, 10, 84486-84517.

https://doi.org/10.1109/ACCESS.2022.3197671

Dalina Dumitrescu / Aprendizaje basado en las preferencias humanas: un estudio piloto sobre la percepción

de los estudiantes sobre el uso de la IA y ChatGPT

645

 Sarmah, S. S. (2019). “Artificial Intelligence in Automation”- RESEARCH REVIEW.

International Journal of Multidisciplinary, 04(06).

 Shamima, A., Alshater, M. M., Ammari, A. E., & Hammami, H. (2022). “Artificial

intelligence and machine learning in finance: A bibliometric review”. Research

in International Business and Finance, 61, 101646.

https://doi.org/10.1016/j.ribaf.2022.101646

 Shangyao, Y., & Yu‐Lin, S. (2007). “A time‐space network model for work team

scheduling after a major disaster.” Journal of the Chinese Institute of

Engineers, 30(1), 63-75. https://doi.org/10.1080/02533839.2007.9671231

 Sharma, R. (2021). “Artificial Intelligence in Agriculture: A Review. “ 5th

International Conference on Intelligent Computing and Control Systems

(ICICCS), 937-942, https://doi.org/10.1109/ICICCS51141.2021.9432187

 Shaik, T., Tao, X., Dann, C., Xie, H., Li, Y., & Galligan, L. (2023). “Sentiment

analysis and opinion mining on educational data: A survey.” Natural Language

Processing Journal, 2, 100003. https://doi.org/10.1016/j.nlp.2022.100003

 Siedlecki, Sandra L,(2022) PhD, RN, APRN-CNS, FAAN. “Conducting Interviews

for Qualitative Research Studies.” Clinical Nurse Specialist 36(2):p 78-80, 3/4

2022. | DOI: 10.1097/NUR.0000000000000653

 Stahl, B. S. (2022). “Responsible innovation ecosystems: Ethical implications of

the application of the ecosystem concept to artificial intelligence.” International

Journal of Information Management, 62, 102441.

https://doi.org/10.1016/j.ijinfomgt.2021.102441

 Sung-Hee Jin, Kowloon Im, Mina Yoo, Ido Roll, Kyoungwon Seo- “Supporting

students’ self-regulated learning in online learning using artificial intelligence

applications” International Journal of Educational Technology in Higher

Education (2023). https://doi.org/10.1186/s41239-023-00406-5,

 Tsolakis, N., Zissis, D., Papaefthimiou, S., & Korfiatis, N. (2022). “Towards AI

driven Mellinger,environmental sustainability: an application of automated

logistics in container port terminals.” International Journal of Production

Research, 60(14), 4508-

4528. https://doi.org/10.1080/00207543.2021.1914355.

 Woolf, B., Burleson, W., Arroyo, I., Dragon, T., Cooper, D. G., & Picard, R. W.

(2009). “Affect-aware tutors: recognizing and responding to student effect.”

International Journal of Learning Technology, 4(3/4), 129-164.

https://doi.org/10.1504/IJLT.2009.028804

 Yao, X., Zhou, J., Zhang, J., & Boër, C. R. (2017). “From Intelligent Manufacturing

to Smart Manufacturing for Industry 4.0 Driven by Next Generation Artificial

Intelligence and Further On," 5th International Conference on Enterprise

Systems (ES) pp. 311-318, Beijing, China.

https://doi.org/https://doi.org/10.1109/ES.2017.58.

Interacción y Perspectiva. Revista de Trabajo Social Vol. 14 No3 / octubre-diciembre, 2024

646

 Yu, K. H., Beam, A. L., & Kohane, I. S. (2018). “Artificial intelligence in

healthcare.” Nature Biomedical Engineering, 2(10), 719-731.

https://doi.org/10.1038/s41551-018-0305-z

 Zhang, Z., Ning, H., Shi, F., Farha, F., Xu, Y., Xu, J., Zhang, F., & Raymond Choo,

K. K. (2022). “Artificial intelligence in cyber security: research advances,

challenges, and opportunities.” Artificial Intelligence Review, 55, 1029-1053.

https://doi.org/10.1007/s10462-021-09976-0

Webinars:

 Webinar : Unlocking the Power of AI: How Tools Like ChatGPT Can Make Teaching

Easier and More Effective-HBP, May 2, 2023-

https://hbsp.harvard.edu/webinars/unlocking-the-power-of-

ai/?cid=email%7Cmarketo%7C2023-05-03-webinar-recording-unlocking-the-

power-of-ai%7C1257639%7Cwebinar%7Cwebinar-registrant%7Cwebinar-

recording-

page%7Cmay2023&acctID=8303506&mkt_tok=ODU1LUFUWi0yOTQAAAGLgOQ

wyjF83cUJGm3uSMFE6b3p10gw0n40c0OMJeqjrgo1991YzwiM5o1Wur8LmOTLI0

P70hpDktd9w6TaWakjWevwHooL97i_Et5uv7XCvA

 Webinar-Will AI Replace the Educator- HBP August 10, 2023 -

https://hbsp.harvard.edu/webinars/will-ai-replace-the-

educator/?cid=email%7Cmarketo%7C2023-08-11-webinar-recording-will-ai-

replace-the-educator%7C1257639%7Cwebinar%7Cwebinar-

registrant%7Cwebinar-recording-

page%7Caug2023&acctID=8303506&mkt_tok=ODU1LUFUWi0yOTQAAAGNg97

w5StIIzze_m5AApAmRjw9rgdw8gWDoUfL3Y03xNYSikCoMf_ym63ya35os2_sxHv

u3N6orlHzC5VEMLOf2AqqEBwOYpBaEz0BdWCxeg

 Webinar- Experimenting with AI in the Classroom-Harvard School of Education-

September 6, 2023- https://www.gse.harvard.edu/ideas/education-

now/23/09/experimenting-ai-

classroom?utm_source=EdNowTY&utm_medium=email&utm_campaign=EdNow

090723