Mualliflar

  • Begim Isoyeva
    Lecturer of Journalism and Mass Communications University of Uzbekistan

DOI:

https://doi.org/10.71337/inlibrary.uz.universaljurnal.101734

Kalit so‘zlar:

sun'iy intellekt baholash avtomatlashtirilgan baholash tizimlari mashina o'rganish tabiiy tilni qayta ishlash (NLP) formativ baholash summativ baholash o'quvchilarning fikr-mulohazalari standartlashtirilgan test.

Annotasiya

Sun'iy intellektning (AI) jadal rivojlanishi turli sohalarga, jumladan, ta'limga sezilarli ta'sir ko'rsatdi. Ushbu sohada sun'iy intellektning eng o'zgaruvchan ilovalaridan biri bu avtomatlashtirilgan baholash tizimlarini ishlab chiqishdir. Ushbu tizimlar talabalar faoliyatini samarali, aniq va adolatli baholash uchun mashinani o'rganish (ML) va tabiiy tilni qayta ishlash (NLP) dan foydalanadi. Baholash jarayonini avtomatlashtirish orqali sunʼiy intellektga asoslangan tizimlar oʻqituvchilarning ish yukini yengillashtiradi, oʻquvchilarga oʻz vaqtida fikr-mulohazalarini taqdim etadi va umumiy baholash jarayonini yaxshilaydi. Bundan tashqari, ular o'rganish naqshlarini aniqlash va shaxsiy tavsiyalar berish uchun keng ko'lamli ma'lumotlarni tahlil qilish imkonini beradi. Biroq, afzalliklariga qaramay, sun'iy intellektga asoslangan baholash tizimlari bir qator qiyinchiliklarga duch keladi, jumladan, AI algoritmlaridagi noto'g'riliklar, yangi pedagogik usullarga moslashmaslik va ma'lumotlarning maxfiyligi va adolatliligi bilan bog'liq axloqiy tashvishlar. Ushbu maqolada AIning avtomatlashtirilgan baholashdagi o‘rni, uning afzalliklari, cheklovlari va axloqiy jihatlari o‘rganiladi va ta’lim natijalarini yaxshilash uchun AIni ham formativ, ham summativ baholashga integratsiyalash strategiyalari muhokama qilinadi.


background image

SUN'IY INTELLEKTNI PEDAGOGIK TA'LIMGA INTEGRATSIYA

QILISH:MUAMMO VA YECHIMLAR mavzusida xalqaro ilmiy-amaliy

anjuman materiallari. Andijon. 23-25 aprel 2025-yil

adpi.uz

universaljurnal.uz

78

AI IN ASSESSMENT: AUTOMATED GRADING SYSTEMS, THEIR

LIMITATIONS AND INTEGRATION INTO FORMATIVE AND SUMMATIVE

ASSESSMENT

Isoyeva Begim

Lecturer of Journalism and Mass

Communications University

Doi:

https://doi.org/10.5281/zenodo.15351396

of Uzbekistan

Abstract:

The rapid advancement of artificial intelligence (AI) has significantly impacted

various fields, including education. One of the most transformative applications of AI in this
domain is the development of automated grading systems. These systems utilize machine learning
(ML) and natural language processing (NLP) to assess student performance efficiently, accurately,
and fairly. By automating the grading process, AI-driven systems alleviate the workload of
educators, provide timely feedback to students, and enhance the overall assessment process.
Additionally, they enable large-scale data analysis to identify learning patterns and offer
personalized recommendations. However, despite their advantages, AI-based grading systems face
several challenges, including biases in AI algorithms, lack of adaptability to new pedagogical
methods, and ethical concerns related to data privacy and fairness. This paper explores the role of
AI in automated grading, its benefits, limitations, and ethical considerations, and discusses
strategies for integrating AI into both formative and summative assessments to improve
educational outcomes.

Keywords:

artificial intelligence, assessment, automated grading systems, machine

learning, natural language pocessing (NLP), formative assessment, summative assessment, student
feedback, standardized testing.

In recent years, artificial intelligence (AI) has revolutionized various sectors, including

education. One of the most significant advancements in this field is the development of automated
grading systems. These systems leverage AI technologies such as machine learning (ML) and
natural language processing (NLP) to assess student performance with greater efficiency,
accuracy, and fairness.[3] By automating the grading process, these systems not only reduce the
workload of educators but also provide students with timely and constructive feedback.
Furthermore, AI-driven grading systems can analyze large volumes of data to identify learning
patterns, detect common errors, and offer personalized recommendations for improvement. As
educational institutions continue to integrate AI into their assessment frameworks, these
technologies have the potential to enhance learning outcomes, bridge educational gaps, and create
a more inclusive and adaptive academic environment. Despite some challenges, the continuous
development of AI-based grading methods holds promise for transforming traditional assessment
strategies and improving the overall educational experience. [5]

Automated grading systems are designed to evaluate different types of assessments,

including multiple-choice questions, essays, and even coding assignments. They analyze student
responses based on predefined criteria and provide immediate feedback, reducing the burden on
educators while enhancing the learning experience for students. As AI continues to evolve, its
ability to assess complex student inputs is improving, making it a viable tool for large-scale
education systems. These systems not only expedite the grading process but also standardize
evaluation criteria, thereby minimizing human inconsistencies and biases. [7] Moreover,
automated grading allows educators to spend more time on personalized instruction and
mentoring, rather than spending long hours reviewing assignments manually.

1. Grading Multiple-Choice and Objective Tests.

AI-powered grading systems efficiently

handle multiple-choice and true/false questions using optical mark recognition (OMR) and rule-
based algorithms. These systems ensure quick and error-free evaluation, allowing educators to


background image

SUN'IY INTELLEKTNI PEDAGOGIK TA'LIMGA INTEGRATSIYA

QILISH:MUAMMO VA YECHIMLAR mavzusida xalqaro ilmiy-amaliy

anjuman materiallari. Andijon. 23-25 aprel 2025-yil

adpi.uz

universaljurnal.uz

79

focus on more complex tasks. Traditional grading methods can be time-consuming, especially in
large classrooms or standardized testing scenarios, but AI-powered tools provide instant results,
enabling students to receive immediate feedback. The implementation of AI in such assessments
also allows institutions to generate in-depth analytics regarding student performance, helping
educators identify knowledge gaps and improve curriculum design. [8]

2. Evaluating Written Responses and Essays.

Natural language processing (NLP) enables

AI to assess student essays and written responses. These systems analyze grammar, coherence,
argument structure, and relevance to the given topic. Machine learning models are trained on vast
datasets to improve their accuracy in evaluating different writing styles and levels. Unlike
traditional grading methods, AI-based grading can provide more detailed feedback, pointing out
specific errors and suggesting improvements in structure, vocabulary, and clarity. Additionally,
AI-powered tools can compare students' responses against large corpora of high-quality essays,
helping to benchmark writing quality and encourage better academic standards. However, while
AI has made significant strides in essay grading, it still faces challenges in evaluating creativity,
originality, and nuanced arguments, necessitating human oversight in some cases. [8]

3. Assessing Coding and Programming Assignments.

Automated grading is also

transforming the evaluation of coding assignments. AI-powered systems check code correctness,
efficiency, and adherence to best practices. They provide instant feedback, allowing students to
improve their skills in real-time. These grading systems utilize techniques such as static code
analysis, dynamic execution testing, and plagiarism detection to ensure that students develop
unique and high-quality solutions. Furthermore, AI-based grading tools can offer step-by-step
debugging suggestions, fostering a more interactive and engaging learning process. By automating
this aspect of assessment, educators can focus on teaching advanced problem-solving skills rather
than spending excessive time reviewing individual lines of code. [8]

Benefits of Automated Grading Systems.

The adoption of AI in assessment brings several

advantages to both educators and students [6]:

·

Efficiency and Speed

: AI systems can evaluate thousands of submissions in minutes,

reducing grading time significantly. This rapid turnaround helps students receive prompt feedback,
allowing them to improve their understanding before progressing to more advanced topics.

·

Consistency and Objectivity

: Unlike human graders, AI provides consistent and unbiased

assessments, minimizing grading discrepancies. Traditional grading can be influenced by
subjective factors such as fatigue, personal bias, or variation in evaluation criteria among different
instructors. AI-based grading eliminates these inconsistencies, ensuring a fair assessment for all
students.

·

Personalized Feedback

: AI can provide detailed insights into student performance,

helping them understand their strengths and areas for improvement. By analyzing previous
submissions, AI can adapt its feedback to individual learners, making recommendations that cater
to specific learning styles.

·

Scalability

: These systems can handle large-scale assessments, making them ideal for

online courses and standardized testing. As education becomes more digitized, AI grading tools
offer a solution for managing vast numbers of students without compromising assessment quality.

·

Cost-Effectiveness

: Automated grading reduces the need for extensive manual grading

resources, allowing institutions to allocate human resources more efficiently.

Challenges and Limitations.

Despite their numerous benefits, automated grading systems

are not without challenges:

·

Understanding Context and Creativity

: AI struggles to accurately assess creative or

context-heavy responses that require human judgment. Essays, research papers, and open-ended
responses often require deep contextual understanding, which AI systems are still working to
master. Unlike human graders, AI lacks the ability to interpret nuances, humor, or abstract
reasoning effectively. Additionally, AI can sometimes misinterpret rhetorical devices and
figurative language, leading to inaccuracies in grading.


background image

SUN'IY INTELLEKTNI PEDAGOGIK TA'LIMGA INTEGRATSIYA

QILISH:MUAMMO VA YECHIMLAR mavzusida xalqaro ilmiy-amaliy

anjuman materiallari. Andijon. 23-25 aprel 2025-yil

adpi.uz

universaljurnal.uz

80

·

Bias in AI Models

: If trained on biased datasets, AI grading systems may reinforce

existing prejudices, affecting fair assessment. For example, AI models that rely on historical data
may inadvertently favor certain linguistic styles or dialects over others. Addressing bias requires
diverse training datasets and continuous evaluation of AI models to ensure equitable grading
outcomes for students from different backgrounds.

·

Technical Limitations

: Complex algorithms require continuous updates and refinements

to improve accuracy and reliability. Maintaining and upgrading AI grading systems demands
ongoing investment and expertise, which can be a barrier for smaller institutions. Additionally, AI
struggles with interpreting handwritten responses, non-standardized formats, and responses that
do not conform to expected patterns.

·

Student and Educator Trust

: Many educators and students remain skeptical about AI’s

ability to fairly and effectively evaluate performance. Concerns include the lack of transparency
in how AI assigns grades, potential errors in assessment, and the possibility of over-reliance on
automation. Establishing transparency in AI grading mechanisms and incorporating human
moderation where necessary can help build trust in these technologies.

·

Ethical Considerations

: The use of AI in grading raises ethical questions about data

privacy, security, and accountability. AI systems store and process vast amounts of student data,
which could be vulnerable to breaches or misuse. Ensuring robust data protection measures and
compliance with ethical guidelines is crucial to maintaining trust in AI-powered assessment tools.
[7]

·

Lack of Adaptability to Evolving Educational Needs

: AI grading systems are typically

designed based on pre-existing grading rubrics and may not easily adapt to new pedagogical
approaches. Traditional AI models rely on rigid evaluation criteria that may not account for the
evolving nature of education, where competency-based and experiential learning approaches are
gaining traction. Future AI systems must incorporate adaptive learning mechanisms to align with
emerging educational methodologies. [7]

The Future of AI in Assessment and Integration into Formative and Summative

Assessment.

The future of automated grading lies in the integration of AI with human expertise.

Hybrid grading models, where AI performs the initial assessment and human graders review
ambiguous cases, can enhance the effectiveness of these systems. Additionally, the advancement
of explainable AI (XAI) is expected to improve transparency in AI grading, allowing students and
educators to understand how AI arrives at its evaluations. AI-driven tools are also expected to
incorporate adaptive learning techniques, where assessments are tailored to each student’s learning
progress, making formative assessments more dynamic and personalized. [1]

AI can be integrated into both formative and summative assessments using tools such as AI-

powered quizzes, real-time feedback mechanisms, and intelligent tutoring systems. For formative
assessments, AI can continuously evaluate student progress through interactive platforms that
provide instant feedback and personalized learning pathways. Summative assessments can benefit
from AI-driven plagiarism detection, automated rubric-based grading, and proctoring technologies
that ensure academic integrity. [4]

Another promising area of development is the integration of AI with blockchain technology,

ensuring the security, transparency, and immutability of grading records. Furthermore, future AI
grading systems will likely incorporate multimodal learning assessments, analyzing not just text-
based responses but also spoken and visual inputs to provide a holistic evaluation of student
competencies.

In conclusion, AI-powered automated grading systems offer a promising solution to modern

educational assessment challenges. By balancing technological advancements with ethical
considerations, these systems can revolutionize learning and assessment methodologies, benefiting
educators and students alike. [8] As AI continues to develop, its potential to reshape the
educational landscape remains immense, providing more efficient, equitable, and insightful ways
to evaluate student performance. However, the integration of AI in grading must be approached


background image

SUN'IY INTELLEKTNI PEDAGOGIK TA'LIMGA INTEGRATSIYA

QILISH:MUAMMO VA YECHIMLAR mavzusida xalqaro ilmiy-amaliy

anjuman materiallari. Andijon. 23-25 aprel 2025-yil

adpi.uz

universaljurnal.uz

81

with caution, ensuring that human oversight and ethical considerations remain central to its
implementation.

References:

1.

Isoeva, B. A. (2024) DIALOGIC TEACHING AS A FORMATIVE ASSESSMENT

TOOL. International Conference on Advance Research in Humanities, Sciences and Education.

Confrencea

,

1

(1), 42–44.

2.

Keynes, S.

(2024).

Transcript: Rethinking the AI Boom, with Daron Acemo

ğ

lu

.

Retrieved from Financial Times

3.

Khasanova, G. (2023). Problem-based learning technology. Journal of Pedagogical

Inventions and Practices, 19, 137-139.

4.

Khasanova, G. K. (2023). ASSESSMENT CRITERIA OF ORGANIZATIONAL-

MANAGERIAL COMPETENCES OF MASTER'S STUDENTS. Oriental renaissance:
Innovative, educational, natural and social sciences, 3(22), 24-29.

5.

LifeWire.

(2024).

The Surprising Ways AI Is Being Used in Schools Right Now

.

Retrieved from LifeWire

6.

National Council of Teachers of English (NCTE).

(2013).

NCTE Position Statement

on Machine Scoring

. Retrieved from NCTE

7.

Perelman, L.

(2012).

Construct Validity, Length, Score, and Time in Holistically

Graded Writing Assessments: The Case against Automated Essay Scoring (AES)

. Retrieved from

WAC Clearinghouse

8.

The Guardian.

(2024).

Voice from the Past: How One University Is Countering AI with

Ancient Examination Techniques

. Retrieved from The Guardian

9.

Wikipedia Contributors.

(2023).

Automated Essay Scoring

. In

Wikipedia, The Free

Encyclopedia

. Retrieved from Wikipedia

Bibliografik manbalar

References:

Isoeva, B. A. (2024) DIALOGIC TEACHING AS A FORMATIVE ASSESSMENT TOOL. International Conference on Advance Research in Humanities, Sciences and Education. Confrencea, 1(1), 42–44.

Keynes, S. (2024). Transcript: Rethinking the AI Boom, with Daron Acemoğlu. Retrieved from Financial Times

Khasanova, G. (2023). Problem-based learning technology. Journal of Pedagogical Inventions and Practices, 19, 137-139.

Khasanova, G. K. (2023). ASSESSMENT CRITERIA OF ORGANIZATIONAL-MANAGERIAL COMPETENCES OF MASTER'S STUDENTS. Oriental renaissance: Innovative, educational, natural and social sciences, 3(22), 24-29.

LifeWire. (2024). The Surprising Ways AI Is Being Used in Schools Right Now. Retrieved from LifeWire

National Council of Teachers of English (NCTE). (2013). NCTE Position Statement on Machine Scoring. Retrieved from NCTE

Perelman, L. (2012). Construct Validity, Length, Score, and Time in Holistically Graded Writing Assessments: The Case against Automated Essay Scoring (AES). Retrieved from WAC Clearinghouse

The Guardian. (2024). Voice from the Past: How One University Is Countering AI with Ancient Examination Techniques. Retrieved from The Guardian

Wikipedia Contributors. (2023). Automated Essay Scoring. In Wikipedia, The Free Encyclopedia. Retrieved from Wikipedia