SCRIPT-SAGE: An AI Chatbot that Helps You Learn Programming

Parth Wani, Aditya Dhonde, Aryabrat Pattanaik, Yash Ghuge and Vaishali Wangikar

MIT Academy of Engineering, Alandi, Pune, India

Keywords:

Artiﬁcial Intelligence (AI) Chatbot, Coding Education, Programming Skills, Computational Thinking,

Learning Modules

Abstract:

This study presents SCRIPT-SAGE, an AI enhanced chatbot aimed at improving programming education,

primirarily in C language. SCRIPT-SAGE provides an engaging and interactive learning experience by com-

bining key features: a chatbot copilot for instant guidance, an integrated code editor for real-time practice,

and quizzes to assess and reinforce understanding. The platform aims to simplify complex programming con-

cepts, promote computational thinking, and support systematic learning through well-structured modules. By

bridging the gap between theoretical knowledge and hands-on coding, SCRIPT-SAGE empowers learners to

develop conﬁdence and proﬁciency in programming in a user-friendly and accessible environment.

1 INTRODUCTION

The emergence of AI has profoundly altered tech-

nological interactions, with chatbots being a promi-

nent advancement in this domain. Chatbots have con-

tributed to a wide range of industries, offering efﬁ-

cient, automated solutions for tasks that once required

human intervention. From customer service to health-

care, e-commerce to ﬁnance, chatbots have made pro-

cesses faster, more accessible, and user-friendly.

Chatbots are now being used in various domains:

• Customer Support: Automating responses to fre-

quently asked questions and troubleshooting is-

sues in real-time.

• Healthcare: Assisting patients with symptom

checks, scheduling appointments, and providing

health-related information.

• E-commerce: Enhancing shopping experiences

by recommending products, guiding customers

through purchasing, and handling order queries.

• Banking and Finance: Facilitating transactions,

helping users track expenses, and offering ﬁnan-

cial advice.

Nevertheless, one of the most consequential ad-

vancements attributed to chatbots resides within the

domain of education, where they are fundamentally

transforming the educational experience. From a so-

ciocultural standpoint, chatbots seem to augment stu-

dents’ social presence by promoting emotional, trans-

parent, and cohesive discourse (P

erez et al., 2020).

Chatbots possess the capacity to provide uniform in-

formation to learners in real-time, incorporating criti-

cal elements such as evaluation standards, submission

deadlines, and the geographical sites of suggested

materials (Cunningham-Nelson et al., 2019). From

pedagogical perspective, unlike traditional methods

that rely on static resources and limited availability

of teachers, chatbots provide ﬂexibility in use, and

are inexhaustibly available. Chatbots demonstrate the

capacity to reiterate material multiple times, thereby

supporting repetitive learning practices. Moreover,

chatbots that integrate both textual and speech-based

functionalities offer valuable assistance to students in

developing comprehensive reading, writing, and lis-

tening skills (Roos, 2018). From a technical stand-

point, the advent of transformer architecture has pro-

foundly transformed the domain through the imple-

mentation of attention mechanisms. These mech-

anisms facilitate large language models (LLMs) in

discerning contextual relationships within sentences

and throughout paragraphs, thus signiﬁcantly enhanc-

ing their efﬁcacy in tasks including translation, sum-

marization, and question-answering. (Abedi et al.,

2023).

This paper provides a comprehensive overview of

the SCRIPT-SAGE project, detailing its system archi-

tecture, core functionalities, and the rationale behind

its design choices. We explore how SCRIPT-SAGE’s

intuitive user interface, progress tracking capabilities,

Wani, P., Dhonde, A., Pattanaik, A., Ghuge, Y. and Wangikar, V.

SCRIPT-SAGE: An AI Chatbot That Helps You Learn Programming.

DOI: 10.5220/0013591400004664

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

In Proceedings of the 3rd International Conference on Futuristic Technology (INCOFT 2025) - Volume 2, pages 301-308

ISBN: 978-989-758-763-4

301

and AI-powered interactions come together to create

a holistic learning environment. By examining the

intended outcomes and potential impact of SCRIPT-

SAGE, we aim to demonstrate its signiﬁcance in the

ﬁeld of computer science education and its potential

to make programming more accessible to a wider au-

dience. SCRIPT-SAGE, an AI-powered application,

leverages these advancements by integrating a built-

in chatbot speciﬁcally designed to help users learn the

C programming language. Through interactive con-

versations, SCRIPT-SAGE guides learners through

complex C programming concepts, offering real-time

code feedback, debugging support, and simpliﬁed ex-

planations to ensure that even beginners can navigate

the intricacies of the language.

2 LITERATURE SURVEY

2.1 A Brief History of Chatbots

A literature review on current chatbot technologies

by (Caldarini et al., 2022) outlines the various types

of chatbots, including the Rule Based chatbots devel-

oped in the primary stages, and the most recenlty used

Deep Learning (DL) based chatbots, the databases

used to train the chatbots, the evaluation metrics used

for each type of chatbot, and their applications in

education and research. Techincal details, history,

and general architecture of chatbots are highlighted

in (Adamopoulou and Moussiades, 2020). Following

are the outline points of a generalized chatbot archi-

tecture:

1. User interface component

2. User message analysis component - A spell-

checker sentiment analysis and machine transla-

tion (multilingual chatbots).

3. Dialog management component - Data handling,

error handling, and ambiguity handling.

2.2 Chatbots in Different Domains

The growing trend in healthcare chatbot develop-

ment addresses the increasing demand for convenient,

at-home medical advice for common ailments like

colds, headaches, and abdominal pain. Research in-

dicates that 60% of doctor visits are for minor condi-

tions, 80% of which can be treated with simple home

remedies, often without needing professional medi-

cal intervention. While existing healthcare chatbots

can provide general advice through frequently asked

questions (FAQs), they lack the ability to offer the

nuanced, natural interaction expected in a human-

doctor consultation. (Gumusel, 2024) The banking

and ﬁnance sectors have rapidly embraced voice as-

sistants and chatbots to offer more responsive and ef-

ﬁcient customer service. In order to improve cus-

tomer interactions, a recent study created a web-based

chatbot for online banking that makes use of artiﬁ-

cial intelligence (AI) capabilities, including natural

language understanding (NLU). This chatbot enables

users to seamlessly access personal banking informa-

tion within the platform, improving the overall bank-

ing experience. (Doherty and Curran, 2019) Chat-

bot technology, often referred to as virtual assistants,

has gained signiﬁcant prominence in the hospitality

industry.The viewpoints of users are not well under-

stood, despite the fact that user perspectives have been

thoroughly examined. A study that addressed this

knowledge gap examined the use of chatbots in the

hotel business using semi-structured interviews with

professionals in the ﬁeld. The results emphasise the

main advantages chatbots offer the hotel sector, such

as enhanced visitor experiences and more efﬁcient op-

erations (Buhalis and Cheng, 2020).

2.3 Chatbots in Education

The beneﬁts of chatbots in the educational system are

discussed by (Adamopoulou and Moussiades, 2020),

noting that the learning support chatbots can preserve

information by replaying previous lessons for stu-

dents who miss them. The use of personal data by

conversational chatbots is covered in (Caldarini et al.,

2022). A security issue arises when chatbots use user

data and even ”learn” from it. Along with how the

backend manages user data on the server, the paper

also discusses security concerns related to data trans-

fers to the chatbot’s server. Some security-related top-

ics discussed in the paper include:

• Authentication and authorization - Malicious

chat-bots operating on crossplatforms, hazards of

sharing data on the internet.

• End-to-End Encryption - Usage of https proto-

col and public-key encryption is encourages.

• Self-destructing messages - Messages contain-

ing sensitive PII (Personally identiﬁable informa-

tion) are automatically erased after a set period of

time.

2.4 Computational Thinking in

Educational Landscape

1. The article (Papadakis, 2022), highlights the need

for developing appropriate applications. Compu-

INCOFT 2025 - International Conference on Futuristic Technology

302

tational Thinking is one of the fundamental skill

of 21st century. Despite the abundance of avail-

able apps, researchers have identiﬁed a critical

need for developmentally appropriate applications

speciﬁcally designed to promote CT concepts and

coding skills in young learners.

2. The literature (Lye and Koh, 2014) emphasizes

three key dimensions of CT: computational con-

cepts, practices, and perspectives. This review

underscores the need for more targeted research

in K-12 settings and highlights the importance of

developing comprehensive instructional strategies

to effectively integrate CT across various educa-

tional contexts.

3. The literature (Fagerlund et al., 2021) reveals a

signiﬁcant challenge in deﬁning clear educational

objectives for CT at the primary level, with cur-

ricula across various countries focusing on related

but distinct areas such as computer science, com-

puting, programming, or digital literacy. This lack

of speciﬁcity extends to the concrete operational-

ization of CT teaching, learning, and assessment

methods, even when using popular programming

environments like Scratch.

2.5 Pedagogical Approaches: Chatbot

Case Study

1. A comprehensive study (Kim and Ko, 2017) an-

alyzing 30 diverse and widely-used online cod-

ing tutorials against dimensions derived from

learning sciences and education literature reveals

both strengths and weaknesses in their approach.

The research indicates that most tutorials share

commonalities in content coverage and employ

a bottom-up organizational structure, providing

goal-directed practice opportunities with immedi-

ate feedback. (Labadze et al., 2023) However, the

study highlights critical shortcomings, notably the

lack of personalization to learners’ prior coding

knowledge and insufﬁcient guidance on knowl-

edge transfer and application.

2. The literature (Bers, 2019) highlights a signiﬁ-

cant gap in pedagogical approaches speciﬁcally

tailored for teaching computer science to young

children. Traditional STEM-based instructional

methods, originally designed for older students,

are often inadequate for early childhood educa-

tion. (Peteranetz et al., 2018) Researchers have

come up with creative solutions to this prob-

lem, like ”Coding as Another Language” (CAL),

which views coding as a new symbolic repre-

sentation system for expressive and communica-

tive purposes. This method, which is based

on principles of language and literacy education,

stresses play, exploration, socialization, and cre-

ativity while acknowledging the distinct devel-

opmental phases of young children. The lit-

erature also introduces the concept of six cod-

ing stages or learning trajectories that children

progress through when exposed to CAL curricu-

lum. Case studies utilizing tools like the KIBO

robot and ScratchJr app have been conducted

to characterize these stages and demonstrate the

practical application of CAL instructional prac-

tices.

3. Study in (Scott et al., 2007) have identiﬁed several

key issues that contribute to student demotivation,

including difﬁculties with programming language

syntax and semantics, underdeveloped problem-

solving and program design skills, and the com-

plexity of development environments. (Popat and

Starkey, 2019) In response to these challenges,

researchers have explored various approaches to

support novice programmers. One such inno-

vative approach involves the use of interactive,

visual problem-solving tools. This paper con-

tributes to this body of research by proposing an

interactive ﬂowchart-based tool that not only aids

in visual problem-solving but also generates syn-

tactically correct program code. The literature

emphasizes the importance of visual representa-

tions in enhancing student understanding, and this

tool builds on that principle by providing ani-

mation features and establishing a clear connec-

tion between visual solutions and code represen-

tations.

3 PROPOSED METHODOLOGY

To address the challenges of teaching programming

through chatbots, we propose SCRIPT-SAGE, an AI-

driven chatbot designed to facilitate coding education.

SCRIPT-SAGE offers an integrated platform where

students can engage in interactive coding lessons, re-

ceive real-time feedback, and improve their coding

skills in a supportive environment. By leveraging AI,

SCRIPT-SAGE aims to enhance the learning experi-

ence by providing personalized assistance, answering

questions, and guiding learners through programming

concepts and practical exercises.

3.1 System Architecture

SCRIPT-SAGE is structured to provide comprehen-

sive experience to the users. Layered architecture in

SCRIPT-SAGE: An AI Chatbot That Helps You Learn Programming

303

Figure 1: Script-Sage: Layered Architecture.

Figure 1 pans out the components of the system: Cod-

ing Playgound, Chatbot, and Quiz Section. User di-

rectly interacts with the top layer consisting of Code

Editor, Code Evaluator, Prompts, Inputs, Questions

and Reviews. All of the components interact with the

core LLM - Gemini by Google. The details of the

interaction are shown in Figure 2. Each data point

generated by user is passed through LLM to convert

the metric data into natural language. This approach

condenses the data and provides user the compact in-

formation. User directly interacts with Code Engine,

Quiz Engine and Chatbot Engine which interacts with

respective utility engines which, in-turn interacts with

Gemini LLM. The approach allows only required in-

formation to ﬂow into the LLM, get the information

from LLM and send it back to the user.

3.2 Technical Aspects

The underlying system mechanism of SCRIPT-SAGE

includes a transformer based chatbot that works based

on attention mechanism. (Vaswani, 2017)

• Self Attention Mechanism: An essential part of

transformer models is the Self-Attention Mecha-

nism. It is essential for comprehending the con-

nections between words in a sentence, which en-

ables the model to successfully capture contextual

meanings. In a sentence, each word’s meaning

can depend on other words in the sentence. Self-

attention calculates how much each word should

pay attention to every other word, helping the

model understand their contextual relationships.

The mechanism computes scores that indicate the

importance of each word in relation to the cur-

rent word being processed. These scores help the

model focus on the relevant words while generat-

ing a response.

• Multiple layers of Attention: Transformers can

recognise the relationships between words in a

sentence regardless of where they are in the sen-

tence thanks to layered layers of self-attention

processes. Each layer reﬁnes the understand-

ing of word connections, allowing the model

to interpret complex dependencies in language.

This layered approach helps the model build a

more detailed representation of the input text.

Unlike older architectures like Recurrent Neural

Networks (RNNs) or Long Short-Term Memory

(LSTM) networks, which process words sequen-

tially, transformers handle all tokens (words or

subwords) in parallel. This parallelism greatly im-

proves the speed of computation, making trans-

formers much more efﬁcient and suitable for pro-

cessing large datasets. It enables the model to bet-

ter manage long-range dependencies.

• Cross Attention Mechanism: The Cross-

Attention Mechanism in multimodal models al-

lows the model to process and relate informa-

tion from different data types (modalities), such

as text, images, audio, and video, to create a uni-

ﬁed understanding of the data. Cross-attention

links information from one data type to another,

allowing the model to interpret the relationships

between them. For example, when processing an

image with a caption, the model can relate speciﬁc

words to particular parts of the image.

Pseudocode for Attention based mechanisms:

1. Tokenization: Convert the input text into tokens.

2. Embedding: Map tokens to vectors using an em-

bedding matrix.

3. For each token in the sequence:

(a) Apply Self-Attention mechanism:

i. Calculate Query, Key, and Value matrices.

ii. Compute attention scores.

iii. Generate a weighted representation of the to-

kens.

(b) Apply Feed-Forward Neural Network to pro-

cess the attended output.

ization.

4. Repeat step 4 for each layer in the transformer

(stacked layers).

5. Generate the output token probabilities using a

Softmax function.

6. Use Beam Search or Sampling to generate the

most probable response.

7. Convert tokens back to human-readable text.

8. Output: Return the generated response to the user.

Query, Key, Value: These matrices are derived

from the input tokens and help in computing the at-

tention scores.

Attention Calculation: This step determines

which words in the sentence should be focused on

more heavily to generate the response.

INCOFT 2025 - International Conference on Futuristic Technology

304

Figure 2: Script-Sage: Working Flow Diagram (DFD).

Softmax Function: Converts raw scores into

probabilities to select the best token for the response.

Feed-Forward Neural Network: Applies trans-

formations to the attended output for deeper process-

ing.

Beam Search/Sampling: Used to generate a co-

herent response by choosing the best possible se-

quence of words.

3.3 UI and Features

Script Sage is designed as an intelligent and inter-

active platform aimed at teaching programming lan-

guages, speciﬁcally focusing on C. Its architecture

revolves around integrating AI-driven chatbot func-

tionalities, real-time coding environments, and user

progress tracking to create a seamless learning expe-

rience. (Tlili et al., 2023)

3.3.1 UI

1. The core of SCRIPT-SAGE in Figure 3, is Google

Gemini 1.5 pro latest, which is tuned to answer

user queries related to coding. The chatbot is

trained on a large corpus of programming knowl-

edge. By providing clear explanations and step-

by-step guidance, it enhances the learning process

for both beginners and intermediate users.

2. SCRIPT-SAGE integrates a real-time code editor

as shown in 4 that supports C programming. This

editor allows users to write and run code directly

within the chatbot interface, offering immediate

Figure 3: Script-Sage: Home.

Figure 4: Script-Sage: Code Editor.

feedback on syntax and logic errors. It also in-

cludes features like syntax highlighting and error

detection, which enhance the learning experience

by providing useful insights into the coding pro-

cess.

3. SCRIPT-SAGE includes interactive quizzes that

evaluate the user’s grasp of programming con-

cepts. The learning module is adaptive, which

SCRIPT-SAGE: An AI Chatbot That Helps You Learn Programming

305

Figure 5: Script-Sage: Question Generator.

Figure 6: Script-Sage: Quiz.

means that it presents content tailored to the user’s

skill level and progress. As users complete the

quizzes (Figure 5), the system generates feedback,

helping them focus on areas that need improve-

ment.

4. A key feature of SCRIPT-SAGE is its ability to

track user progress over time. The system records

the modules completed, quiz scores , and cod-

ing exercises, allowing users to monitor their im-

provement. This data can also be used to person-

alize future learning materials.

3.3.2 Features

Key components of the system include the following:

1. The core of Script Sage is a generative AI chatbot

that helps users learn programming by answering

questions, explaining code, and debugging prob-

lems. The chatbot leverages a large dataset of

programming-related content, enabling it to of-

fer context-aware responses based on the user’s

queries.

2. The system integrates a code editor within the in-

terface where users can write and execute C pro-

gramming code. This editor provides immediate

feedback on syntax errors, logic issues, and per-

formance, making the learning process more prac-

tical and hands-on.

3. A dynamic learning module that adapts to the

user’s knowledge level, providing interactive

quizzes to assess and reinforce their understand-

ing of programming concepts. Based on perfor-

mance, the module adjusts the difﬁculty of subse-

quent questions and provides detailed feedback.

4. Script Sage tracks user activities, including quiz

scores, completed modules, and coding time

spent. This data is analyzed to create personalized

learning paths, ensuring that users focus on areas

that require improvement while encouraging con-

tinuous engagement.

5. The system incorporates elements of casual con-

versation designed to support the mental well-

being of the user. These include reminders to take

breaks during long coding sessions and motiva-

tional feedback to encourage consistent learning

without causing burnout.

6. The NLP component of Script Sage is crucial for

understanding user input and delivering appropri-

ate responses. The model is ﬁne-tuned for pro-

gramming education, ensuring that the chatbot

can handle queries related to:

(a) Syntax explanation and logic-building.

(b) Debugging and error handling.

The chatbot can interpret user intent, regardless of

the query’s complexity, and provide personalized

assistance. Over time, the model can be improved

using user data to enhance response accuracy and

engagement.

4 COMPARATIVE ANALYSIS

4.1 Detailed Insights

• Zendesk’s Answer Bot (Customer Service):

– Functionality: Focuses on automating cus-

tomer service interactions by providing instant

answers to frequently asked questions.

– Strengths: Signiﬁcantly reduces response

times, enhancing customer satisfaction.

• Woe-bot (Healthcare):

– Functionality: Offers mental health support

using cognitive behavioral therapy techniques.

– Strengths: Provides accessible emotional sup-

port in a friendly manner.

• Duolingo (Education):

INCOFT 2025 - International Conference on Futuristic Technology

306

Table 1: Domains and their corresponding chatbot functionalities. The ﬁnal sentence of a caption should end with a period.

Domain Chatbot Functionality

Programming Script Sage C programming queries, code snippets, debugging

assistance.

Customer Service Zendesk’s Answer Bot Automates responses to customer queries.

Healthcare Woebot Mental health support and CBT techniques.

Education Duolingo Language learning through interactive conversa-

tions.

Finance Cleo Personal ﬁnance management and budgeting ad-

vice.

Entertainment AI Dungeon Interactive storytelling and creative writing.

– Functionality: Engages users in language

learning through games and quizzes.

– Strengths: Its gamiﬁed approach keeps users

motivated and enhances language retention.

• Cleo (Finance):

– Functionality: Acts as a personal ﬁnance as-

sistant, offering insights on spending and bud-

geting.

– Strengths: Provides users with a clear picture

of their ﬁnancial health.

• AI Dungeon (Entertainment):

– Functionality: Allows users to create and navi-

gate through interactive storytelling adventures.

– Strengths: The open-ended nature of interac-

tions fosters creativity.

• SCRIPT-SAGE:

– Specialization: Designed speciﬁcally for C

programming, it excels in providing targeted

support for coding-related queries.

– Strengths: Offers deep insights into C lan-

guage features, including memory management

and data structures.

5 CONCLUSION

The SCRIPT-SAGE project exempliﬁes how modern

technology, particularly AI, can be used to create a

dynamic and engaging learning environment. Us-

ing generative AI to power its conversational chat-

bot, SCRIPT-SAGE provides contextually relevant

personalized support to learners. The integration of

a real-time code editor, quizzes, and user progress

tracking ensures that users receive immediate feed-

back and can monitor their learning journey effec-

tively.

The technology stack, consisting of React JS for

the frontend and AWS cloud services for deployment,

enables scalability, performance, and a seamless user

experience. The use of CI/CD pipelines ensures that

updates to chatbot and learning modules are deployed

efﬁciently, maintaining the quality and availability of

the application.

In future phases, the deployment of SCRIPT-

SAGE in the cloud will allow continuous improve-

ments in system performance, scalability, and user

interaction, laying the groundwork for a long-lasting

and impactful tool in the world of programming edu-

cation.

REFERENCES

Abedi, M., Alshybani, I., Shahadat, M. R. B., and Murillo,

M. (2023). Beyond traditional teaching: The poten-

tial of large language models and chatbots in graduate

engineering education. Qeios.

Adamopoulou, E. and Moussiades, L. (2020). An overview

of chatbot technology. In IFIP international confer-

ence on artiﬁcial intelligence applications and inno-

vations, pages 373–383. Springer.

Bers, M. U. (2019). Coding as another language: A ped-

agogical approach for teaching computer science in

early childhood. Journal of Computers in Education,

6(4):499–528.

Buhalis, D. and Cheng, E. S. Y. (2020). Exploring the use of

chatbots in hotels: technology providers’ perspective.

In Information and Communication Technologies in

Tourism 2020: Proceedings of the International Con-

ference in Surrey, United Kingdom, January 08–10,

2020, pages 231–242. Springer.

Caldarini, G., Jaf, S., and McGarry, K. (2022). A literature

survey of recent advances in chatbots. Information,

13(1):41.

Cunningham-Nelson, S., Boles, W., Trouton, L., and Marg-

erison, E. (2019). A review of chatbots in education:

practical steps forward. In 30th annual conference

for the australasian association for engineering ed-

ucation (AAEE 2019): educators becoming agents of

change: innovate, integrate, motivate, pages 299–306.

Engineers Australia.

SCRIPT-SAGE: An AI Chatbot That Helps You Learn Programming

307

Doherty, D. and Curran, K. (2019). Chatbots for online

banking services. In Web Intelligence, volume 17,

pages 327–342. IOS Press.

Fagerlund, J., H

akkinen, P., Vesisenaho, M., and Viiri,

J. (2021). Computational thinking in programming

with scratch in primary schools: A systematic re-

view. Computer Applications in Engineering Educa-

tion, 29(1):12–28.

Gumusel, E. (2024). A literature review of user privacy con-

cerns in conversational chatbots: A social informatics

approach: An annual review of information science

and technology (arist) paper. Journal of the Associa-

tion for Information Science and Technology.

Kim, A. S. and Ko, A. J. (2017). A pedagogical analy-

sis of online coding tutorials. In Proceedings of the

2017 ACM SIGCSE Technical Symposium on Com-

puter Science Education, pages 321–326.

Labadze, L., Grigolia, M., and Machaidze, L. (2023). Role

of ai chatbots in education: systematic literature re-

view. International Journal of Educational Technol-

ogy in Higher Education, 20(1):56.

Lye, S. Y. and Koh, J. H. L. (2014). Review on teaching and

learning of computational thinking through program-

ming: What is next for k-12? Computers in human

behavior, 41:51–61.

Papadakis, S. (2022). Apps to promote computational

thinking and coding skills to young age children:

A pedagogical challenge for the 21st century learn-

ers. Educational Process: International Journal

(EDUPIJ), 11(1):7–13.

erez, J. Q., Daradoumis, T., and Puig, J. M. M. (2020). Re-

discovering the use of chatbots in education: A sys-

tematic literature review. Computer Applications in

Engineering Education, 28(6):1549–1565.

Peteranetz, M. S., Wang, S., Shell, D. F., Flanigan, A. E.,

and Soh, L.-K. (2018). Examining the impact of com-

putational creativity exercises on college computer

science students’ learning, achievement, self-efﬁcacy,

and creativity. In Proceedings of the 49th ACM techni-

cal symposium on computer science education, pages

155–160.

Popat, S. and Starkey, L. (2019). Learning to code or coding

to learn? a systematic review. Computers & Educa-

tion, 128:365–376.

Roos, S. (2018). Chatbots in education: A passing trend or

a valuable pedagogical tool?

Scott, A., Watkins, M., and McPhee, D. (2007). A step back

from coding–an online environment and pedagogy for

novice programmers. In Proceedings of the 11th Java

in the Internet Curriculum Conference, volume 2007,

pages 35–41. Citeseer.

Tlili, A., Shehata, B., Adarkwah, M. A., Bozkurt, A.,

Hickey, D. T., Huang, R., and Agyemang, B. (2023).

What if the devil is my guardian angel: Chatgpt as

a case study of using chatbots in education. Smart

learning environments, 10(1):15.

Vaswani, A. (2017). Attention is all you need. Advances in

Neural Information Processing Systems.

INCOFT 2025 - International Conference on Futuristic Technology

308