Artificial intelligence in preschool education: Personalized assessment and evaluation with artificial intelligence

161 Gülçin GÜVEN & Seher YUMUGAN USBED, 7(12), 2025, İlkbahar / Spring For citation / Atıf için: GÜVEN, G., & YUMUGAN, S. (2025). Artificial intelligence in preschool education: Personalized assessment and evaluation with artificial intelligence. Uluslararası Sosyal Bilimler ve Eğitim Dergisi – USBED 7(12), 161–190. https://doi.org/10.5281/zenodo.14936787, https://dergipark.org.tr/tr/pub/usbed Artificial intelligence in preschool education: Personalized assessment and evaluation with artificial intelligence Gülçin GÜVEN 1 Assoc. Prof. Dr; Marmara Unıversity, Ataturk Faculty of Education, Early Childhood Education, 34854, İstanbul, Türkiye E-mail: [email protected] ORCID: 0000-0002-9638-025X Seher YUMUGAN Study; Marmara University, Ataturk Faculty of Education, Preschool Teaching, 34854, İstanbul, Türkiye E-mail: yumugans3[email protected] ORCID: 0009-0004 -8809 -5012 Article Type: Research Article Submission Date: 29/01/2025 Revision Dates: 30/01/2024 (Editor c.), 21/02/2025 (Minor r.) Accepted Date: 27/02/2025 Ethics Statement ✓ The author(s) declare(s) that the study is not subject to ethics committee approval. ✓ The study does not require ethical approval. Researchers' contribution to the study 1. Author contribution: Wrote the article, collected the data, and analyzed/reported the results (50%). 2. Author contribution: Wrote the article, collected the data, and analyzed/reported the results (50%). Conflict of interest The author(s) declare(s) that there is no potential conflict of interest in this study. Similarity This study was scanned in the iThenticateprogram. The final similarity rate is 8%. Corresponding author 162 Artificial intelligence in preschool education: Personalized assessment and evaluation with artificial intelligence USBED, 7(12), 2025, İlkbahar / Spring Artificial intelligence in preschool education: Personalized assessment and evaluation with artificial intelligence Abstract Traditional assessment and evaluation methods in preschool education lack objectivity, limit process-oriented assessment, and fail to account for individual differences, reducing their overall effectiveness. This study aims to address these challenges by designing an artificial intelligence-supported model to enhance assessment and evaluation processes. The research examines the integration of assessment tools in Turkey’s Preschool Education Program into an AI-supported model, focusing on the use of natural language processing (NLP), image recognition, and machine learning algorithms for a more detailed and objective evaluation of children's cognitive, social, and emotional development. A qualitative research method was employed, utilizing document and content analysis techniques. A comprehensive literature review was conducted, and the findings were systematically analyzed to develop a new AI-supported assessment model. The proposed model enhances individualized learning experiences, integrates AI into educational processes, and provides real-time feedback to students and teachers. This study presents an innovative alternative to traditional approaches by introducing AI-based assessment in Turkish preschool education. The model is expected to improve objectivity, consistency, and sensitivity to individual differences, offering a more efficient and data-driven evaluation framework for early childhood education. Keywords: Artificial intelligence supported measurement and evaluation, Preschool education, Personalized learning, Data driven feedback, Educational technologies EXTENDED ABSTRACT Introduction Traditional assessment and evaluation methods in early childhood education (ECE) often fall short of providing objective, process-oriented, and individualized evaluations. These limitations restrict the effectiveness of assessment processes and hinder the ability to track the cognitive, social, and emotional development of children holistically. The present study aims to address these challenges by proposing an artificial intelligence (AI)- supported model that integrates innovative technologies, such as natural language processing (NLP), image recognition, and machine learning algorithms, into ECE assessment processes. By leveraging AI-supported personalized assessment models, this study seeks to enhance real-time feedback mechanisms and support individualized learning experiences in early childhood education. The research is significant as it proposes an alternative to traditional assessment methodologies, providing a more reliable, adaptive, and inclusive model. The study contributes to the literature by examining the integration of AI technologies into the Turkish ECE system and evaluating their potential benefits and challenges. The study also identifies various constraints encountered during the research process, including limitations in data accessibility, ethical considerations in AI applications for children, and infrastructural challenges in implementing AI-supported assessment models in real-world educational settings. Conceptual and Theoretical Framework Concepts Key concepts addressed in this study include: • Artificial Intelligence in Education (AIED): AI-driven technologies that enhance educational assessment, learning experiences, and personalized feedback. • Personalized Learning: Tailoring educational experiences to individual students' needs, preferences, and learning speeds through adaptive assessment tools. • Data-Driven Feedback: AI-supported real-time assessment methods that provide educators with datadriven insights into children's cognitive, social, and emotional development. Existing literature highlights the importance of assessment and evaluation in ECE, with studies emphasizing formative assessment methods and alternative assessment tools. Shepard (2000) and Black and Wiliam (2009) suggest that formative assessments are crucial for tracking students’ development dynamically. Additionally, 163 Gülçin GÜVEN & Seher YUMUGAN USBED, 7(12), 2025, İlkbahar / Spring Holmes et al. (2019) argue that AI-based learning analytics and assessment models can enhance learning outcomes and increase the objectivity of assessments. While traditional assessment methods rely heavily on teacher observations and subjective evaluations, AIsupported assessment tools offer systematic and real-time tracking of student progress. International research (e.g., Zawacki-Richter et al., 2019) indicates that AI can provide automated, adaptive, and data-driven evaluations that reduce bias and improve decision-making in educational settings. In Turkey, research on AI applications in education remains limited, particularly in ECE. However, studies by Akdeniz and Özdinç (2021) suggest that AI has the potential to revolutionize assessment methods by offering individualized learning opportunities and improving feedback mechanisms for both students and educators. Method The study employs a qualitative research methodology with a descriptive design to investigate the integration of artificial intelligence in early childhood education (ECE) assessment. This approach enables an in-depth examination of current evaluation practices and the theoretical potential of AI-supported personalized models (Creswell, 2013; Merriam & Tisdell, 2016). Data Collection and Analysis • Data Sources: Data were gathered through document analysis of national and international academic studies, theses, curricula, legislation, and policy documents related to AI applications in ECE. Purposive sampling ensured that only reliable, relevant, and up-to-date documents were included (Bowen, 2009). • Analysis Techniques: Content analysis was utilized to systematically identify key themes and trends in AI-supported assessment methods (Krippendorff, 2013). A three-stage coding process was implemented: - Open Coding: Detailed examination of documents to extract meaningful expressions, concepts, and themes using established theoretical frameworks. - Axial Coding: Grouping of identified themes based on relationships and similarities to form a hierarchical structure. - Selective Coding: Integration of main and sub-themes into a coherent thematic framework through crossvalidation and iterative discussions, enhancing the study’s reliability (Patton, 2015). Through thematic organization, categorization, and interpretative analysis, the study systematically examines both the advantages and challenges of integrating AI into ECE assessment models (Braun & Clarke, 2006; Miles, Huberman, & Saldaña, 2014). Findings The study identifies several limitations of traditional assessment approaches, highlighting key challenges such as subjectivity and bias, time and resource constraints, and the lack of individualization. Teacher-led evaluations often lack objectivity, leading to inconsistencies in assessment results. Additionally, teachers face difficulties in conducting comprehensive assessments due to limited time and resources, which affects the overall quality and reliability of evaluations. Traditional methods also fail to adequately consider students’ individual learning styles and progress rates, limiting their effectiveness in addressing diverse educational needs. To address these challenges, the study proposes an AI-supported assessment model that integrates several key components. Natural language processing (NLP) is utilized to analyze children's verbal and written communication skills, providing deeper insights into language development. Image recognition technologies are applied to observe children's physical and social interactions in educational settings, ensuring a more comprehensive evaluation of their social and motor skills. Machine learning algorithms personalize learning experiences by adapting assessment criteria based on students' performance data, allowing for a more individualized approach. Additionally, adaptive learning platforms provide real-time tracking of student progress and offer individualized feedback to optimize learning outcomes. The implementation of this AI-supported model enhances the accuracy and objectivity of assessment processes, reducing inconsistencies associated with traditional methods. Real-time feedback mechanisms support teachers in making data-driven educational decisions, enabling more informed instructional planning. Adaptive learning platforms facilitate customized educational interventions by addressing individual student needs and learning paces. Furthermore, AI applications help identify learning gaps and provide targeted support to students requiring additional assistance, fostering a more inclusive and effective educational environment. 164 Artificial intelligence in preschool education: Personalized assessment and evaluation with artificial intelligence USBED, 7(12), 2025, İlkbahar / Spring Conclusion, Discussion, and Recommendations The findings of this study align with existing research highlighting the benefits of AI-based assessment models in education. The implementation of AI in early childhood education assessment presents a transformative approach to evaluating children’s development more accurately and objectively. AI-supported assessment models increase objectivity, efficiency, and reliability in educational evaluations. Personalized learning experiences can enhance student engagement and improve educational outcomes. AI applications enable a more holistic and data-driven approach to early childhood education assessment. Despite these advantages, the implementation of AI-based assessment models requires appropriate technological infrastructure and comprehensive teacher training. Ethical concerns, such as data privacy and AI bias, must also be carefully addressed to ensure the responsible use of AI in educational settings. Future research should focus on testing AI-supported assessment models in real-world educational environments to evaluate their long-term impact and effectiveness. To maximize the benefits of AI-supported assessment, educators should be equipped with the necessary skills to effectively use AI in assessment processes through targeted teacher training programs. Additionally, government agencies should establish clear policies and guidelines for the ethical and responsible integration of AI in education. Further empirical studies are necessary to assess the effectiveness and feasibility of AI-supported assessment models across different educational contexts, ensuring their adaptability and sustainability in diverse learning environments. This study provides a groundbreaking perspective on the integration of AI into early childhood education assessment. By leveraging AI technologies, educators can create a more personalized, efficient, and objective assessment environment. The proposed AI-supported model offers a viable alternative to traditional assessment methods, ensuring that the learning experiences of children are accurately evaluated and continuously improved. INTRODUCTION The concept of education has been defined in various ways by different thinkers and educators. The differences in these definitions stem from the fact that individuals have different perspectives on education. According to Senemoğlu (2005), education is seen as "the process of nurturing human personality" and "investment in human capital". Sönmez (2023) defines education as "the process of creating desired biochemical changes in the brain as a result of physical stimulation". In the most general sense, in the words of Ertürk (1972), education is considered as the process of "creating desired behavior" or "changing desired behavior" (Başaran, 2000; Senemoğlu, 2005; Sönmez, 2023). Education is a process that aims to develop human beings, to reveal their talents, to direct them to the good and the beautiful, and to create desired behavioral changes, starting with birth and continuing until death (Şişman, 2021). Preschool education is a systematic educational process that directs all the development of children aged 0-72 months in line with the cultural values of the society; helps them in the reasoning process by increasing the development of their emotions and perception power; develops their creativity, enables them to express themselves and gain self-control (Güven & Azkeskin, 2010). Preschool education is a critical period that supports children's cognitive, emotional and social development processes. In this process, measurement and evaluation play an important role in improving the quality of education and enabling individualized learning experiences. Today, the main objectives of preschool education include monitoring children's 165 Gülçin GÜVEN & Seher YUMUGAN USBED, 7(12), 2025, İlkbahar / Spring development, understanding their individual differences, and organizing education programs accordingly (Bayrak & Duruhan, 2013). In Turkey, preschool education is carried out within the framework of a program prepared in advance as formal education and developed in line with the needs of the age and society (Gelebek Üstün & Uzun, 2020). Early childhood educators believe that the individual interests and needs of the child should be taken into consideration in the preparation of the program and evaluation (Mağden, 1993; Sapsağlam, 2013). The behaviors acquired in this period affect the whole life of the child and therefore a planned education program is essential (Bayrak & Duruhan, 2013). Measurement and evaluation is one of the fundamental elements of the educational process. In the most general sense, evaluation is defined as drawing conclusions from measurements and making a value judgment about the measured individuals or objects (Atılgan, Kan, & Doğan, 2019). While measurement indicates the amount of a characteristic, evaluation indicates whether this amount is sufficient or suitable for the purpose (Küçükahmet, 2009). Evaluation is a broader concept that includes measurement and involves the process of making decisions about the measured characteristics of individuals by comparing the measurement results with a criterion. The knowledge and skills that teachers acquire in the field of measurement and evaluation have a critical importance for their professional lives. However, it is not easy to claim that all teachers have sufficient knowledge and skills in the field of assessment and evaluation (Yaşar, 2014). Especially in the implementation of alternative assessment and evaluation approaches, teachers need to have sufficient knowledge (Yenice, Özden, & Alpak Tunç, 2017). It is very important to follow the development of children in preschool period. Therefore, it is necessary to use various assessment and evaluation tools to monitor and support the development of preschool children (Bayrak & Duruhan, 2013). Measurement and evaluation in preschool period is a broad and comprehensive process for the purpose of recognizing the child and planning education (Sapsağlam, 2013). The evaluation process is not only an action that takes place on paper, but also involves the knowledge, awareness and subjectivity of the person making the evaluation (Bentzen, 2005). However, existing research shows the limitations of traditional assessment and evaluation methods. Gül Can (2021) points out that methods based on observation and teacher evaluations involve subjectivity, which makes it difficult to reflect children's actual developmental processes. Çetin (2020) states that these methods do not sufficiently take into account individual differences and are result-oriented instead of processoriented. 166 Artificial intelligence in preschool education: Personalized assessment and evaluation with artificial intelligence USBED, 7(12), 2025, İlkbahar / Spring These limitations create serious disadvantages especially for children with special needs or children in mainstreaming education. In the literature, the variety of measurement tools used in the preschool period draws attention. In Turkey, the measurement tools recommended by the Ministry of National Education include skill observation forms, development reports, portfolios and anecdotal record forms (MEB, 2013). These tools are used to monitor children's cognitive, social, emotional and physical development. However, it is known that teachers face difficulties in the implementation of these tools, such as lack of time, large class sizes and lack of materials (Ata & Bolat, 2022). In the international literature, the importance of assessment and evaluation in the preschool period has been widely discussed. Researchers such as Shepard (2000) and Black and Wiliam (2009) emphasize that formative assessment methods address children's developmental processes in a more dynamic and flexible way. In addition, artificial intelligence-supported systems aim to support teachers by offering innovative solutions in assessment and evaluation processes (Holmes et al., 2019). More studies need to be conducted on the functionality and applicability of the measurement tools in the current Preschool Education Program in Turkey. The 2013 Preschool Education Program of the Ministry of National Education includes different assessment and evaluation methods to evaluate children. These methods include development observation forms, development reports, portfolios, checklists, anecdotes, and observation records (MEB, 2013; Gelebek Üstün & Uzun, 2020). The 2024 Maarif Program is an approach to preschool education that aims to support children's cognitive, emotional, social and physical development. The program structures educational processes by taking into account children's individual differences to ensure their healthy development at an early age. It also encourages children's active participation in activities and learning processes. By adopting a multidimensional evaluation approach, the program creates a continuous feedback mechanism between the child, the teacher and the program. This approach aims to use a variety of assessment tools to monitor and support children's development and to support teachers' professional development. The 2024 Maarif Program also aims to provide a more effective educational experience through the diversification of educational materials and methods and the integration of innovative technologies. In the Maarif Model 2024 Preschool Education Program, the main purpose of the measurement and evaluation process is to support children's skill acquisition process, monitor their development and increase the effectiveness of teaching processes. Various measurement tools are used to monitor children's cognitive, social, emotional, physical and language development. By 167 Gülçin GÜVEN & Seher YUMUGAN USBED, 7(12), 2025, İlkbahar / Spring adopting a multidimensional evaluation approach, the program creates a continuous feedback mechanism for the child, the teacher and the program itself (MEB, 2024). In this context, the assessment tools used in the program play an important role in evaluating the effectiveness of the program as well as providing continuous feedback to teachers by monitoring children's development. Although traditional assessment and evaluation methods are widely used in education, they have various weaknesses. The most obvious weakness of these methods is their lack of objectivity. Methods based on observation and teacher evaluations can be subjective depending on the individual interpretations of the teacher. This makes it difficult to evaluate each student's development process in a consistent and unbiased manner (Gül Can, 2021). When teachers evaluate children based on their personal perspectives, objective results may not be obtained (Sezer, 2010). Another important problem is that these methods fail to take into account the individual needs and learning styles of all students. Every child has a different learning pace and style, but traditional methods ignore these differences (Çetin, 2020). This creates a major disadvantage, especially for children who are mainstreamed or have special needs (Gül Can, 2021). Traditional methods are also outcome-oriented rather than process-oriented. Instead of following the child throughout the developmental process, the focus is usually on the final results. This makes it difficult to adequately understand the child's developmental stages and can lead to generalizations based only on specific results (Sezer, 2010). While the child's immediate performances are taken into consideration, the learning process and individual progress may be ignored (Çetin, 2020). The importance of performance-based assessment methods increases in this context. In particular, tools such as portfolio assessment offer an effective solution for tracking children's individual learning processes. Portfolios provide a more systematic way of tracking children's learning outcomes, work and development. Similarly, methods such as anecdotal records allow for recording children's individual achievements and behaviors throughout the process. These tools enable a more in-depth analysis of each child's individual developmental needs through a process-oriented assessment. Skill Observation Form: This form is used to observe and report the extent to which children acquire different domain skills (Turkish, math, science, social, movement and health, art, music). Teachers observe children and record their level of skill acquisition. 168 Artificial intelligence in preschool education: Personalized assessment and evaluation with artificial intelligence USBED, 7(12), 2025, İlkbahar / Spring Anecdotal Record Form: Individual achievements, skills or behaviors of children that stand out during their learning process are recorded by teachers. These records are used to understand children's individual development and to provide feedback on the process. Skill Acquisition Report: The skills that children acquire during their developmental periods are reported in line with the criteria set. These reports are important for determining in which areas children need to be supported. Learning Evidence: In line with the learning outcomes in the program, the products (drawings, projects, presentations, etc.) in which children can demonstrate what they have learned are evaluated. Evidence of learning shows how children apply their knowledge and skills and is included in teachers' assessment processes. Monthly and Daily Plans: Monthly and daily education plans prepared by teachers are used to observe and monitor children's skill acquisition throughout the process. These plans systematically evaluate children's progress in their learning experiences. Family Participation Preference Form: Used for families to indicate how they would like to participate in the educational process. This form is important to assess children's learning processes at home and to ensure cooperation between school and home. These tools allow teachers to closely monitor children's development. At the same time, these data are used to evaluate the effectiveness of the program and support teachers' professional development (MEB, 2024). However, research shows that teachers experience various difficulties in using assessment and evaluation tools (Ata & Bolat, 2022). Teachers face obstacles such as lack of time, excessive class size, lack of materials and multiplicity of assessment tools (İnce, 2022). Teachers' lack of knowledge about measurement and evaluation methods reduces the effectiveness of evaluation processes. In addition, it has been found that teachers have difficulty in applying tools such as checklists, anecdotes, observations, and portfolios, and therefore cannot adequately carry out measurement and evaluation processes (Sezer, 2010; Ata & Bolat, 2022). This situation may have negative consequences in terms of monitoring and supporting children's development. Measurement and evaluation processes in the preschool period are critical for identifying children's developmental needs and preparing individualized education plans. As emphasized in the 2024 Education Program of the Ministry of National Education, the diversity and 169 Gülçin GÜVEN & Seher YUMUGAN USBED, 7(12), 2025, İlkbahar / Spring accuracy of the tools used in these processes are seen as the cornerstones to increase the effectiveness of education (MEB, 2024). In addition to observation-based approaches in assessment and evaluation processes, the importance of dynamic and individual developmentoriented methods through the integration of technology is frequently emphasized in the international literature (Shepard, 2000; Holmes et al., 2019). In this direction, it is necessary to overcome the knowledge and skill deficiencies of teachers in order to use assessment tools correctly and to increase the applicability of these tools. In particular, formative assessment methods are reported to effectively support both children's developmental processes and learning outcomes (Black & Wiliam, 2009). Measurement and evaluation processes in the preschool period are critical for identifying children's developmental needs and preparing individualized education plans. As emphasized in the 2024 Education Program of the Ministry of National Education, the diversity and accuracy of the tools used in these processes are seen as the cornerstones to increase the effectiveness of education (MEB, 2024). In addition to observation-based approaches in assessment and evaluation processes, the importance of dynamic and individual developmentoriented methods through the integration of technology is frequently emphasized in the international literature (Shepard, 2000; Holmes et al., 2019). In this direction, it is necessary to overcome the knowledge and skill deficiencies of teachers in order to use assessment tools correctly and to increase the applicability of these tools. In particular, formative assessment methods are reported to effectively support both children's developmental processes and learning outcomes (Black & Wiliam, 2009). Measurement and evaluation in preschool education has a critical importance in terms of determining the individual differences of children, identifying their strengths and weaknesses, and shaping the educational process according to these data. Increasing the knowledge and skills of teachers in the field of measurement and evaluation is necessary to eliminate the difficulties encountered in this process and to implement education programs effectively. Although traditional assessment and evaluation methods are widely used in education, they have various weaknesses. The most obvious weakness of these methods is their lack of objectivity. Methods based on observation and teacher evaluations can be subjective depending on the individual interpretations of the teacher. This makes it difficult to evaluate each student's developmental process in a consistent and unbiased manner (Gül Can, 2021). When teachers evaluate children based on their personal perspectives, objective results may not be obtained (Sezer, 2010). Especially in the preschool period, teachers' subjective evaluations with limited 176 Artificial intelligence in preschool education: Personalized assessment and evaluation with artificial intelligence USBED, 7(12), 2025, İlkbahar / Spring vocabulary, grammar usage and expressive skills can be analyzed in real time. Thus, individualized education plans can be created by identifying a child's strengths and weaknesses in language development. In one study, NLP methods were used to analyze children's storytelling abilities and these analyses were used to suggest to teachers what kind of activities would be more effective in supporting children's language development. Such a practice contributes to a more effective educational process by providing an accurate and objective assessment of children's language skills. For example, in read-aloud activities, NLP models can determine children's learning levels from the words they repeat and which types of words they are more interested in. This data makes it possible to design content and activities that support children's language development more efficiently. In this study, the accuracy of news articles was analyzed using the Long Short Term Memory (LSTM) model with a high success rate, which shows the potential of language analysis-based applications in educational processes (Toğaçar et al., 2022). The use of natural language processing (NLP) technologies in preschool education has significant potential to support children's language development and integrate innovative approaches into educational processes. Language development at an early age is one of the key components of cognitive, emotional and social development. In this context, AI-based NLP models such as LSTM offer an effective tool to detect difficulties and progress in language development by analyzing children's language use, vocabulary and meaning-making processes. NLP technologies can contribute to the development of individualized language learning programs, enabling the design of educational materials and strategies appropriate to each child's language development level. In addition, more efficient and personalized learning experiences can be created by using NLP methods in areas such as evaluating the appropriateness of educational content for children's age and language levels, designing gamified applications to improve language skills, and instant detection of errors in language use. In this framework, artificial intelligencesupported language analysis and personalized feedback mechanisms offer an innovative approach to optimize language teaching in preschool education. Image recognition technologies can be used to monitor children's physical activities and social interactions (Nguyen et al., 2020). In Turkey, research on the use of these technologies in education provides examples of applications and recommendations. Research on image recognition technologies evaluates how these technologies can be applied in education. Object recognition studies using deep learning methods such as Faster R-CNN have shown that these 177 Gülçin GÜVEN & Seher YUMUGAN USBED, 7(12), 2025, İlkbahar / Spring technologies can be used to monitor students' motor skills and social interactions. In this way, children's classroom activities can be monitored and analyzed more closely (Yılmaz, Aydın, & Çetinkaya, 2020). In addition, machine learning algorithms can identify children's learning styles and developmental rates to provide customized educational materials (Luckin et al., 2016). Adaptive learning platforms can enrich the learning experience by supporting each child's individual learning journey (Holmes et al., 2019). Studies on adaptive learning and machine learning in Turkey discuss the potential and applicability of these technologies in education. Studies on machine learning show that these technologies provide significant advantages in determining students' individual learning rates and providing personalized learning experiences. The use of artificial intelligence (AI) technologies in education offers important innovations, especially in areas such as personalized learning, automated assessment and improvement of teaching processes. Various studies in Turkey have examined this issue. For example, Çavuş (2024) examined the benefits of AI-based assessment and evaluation tools in the field of education, the difficulties experienced in the process and the role of teachers in this process. Tonbuloğlu (2023) states that the use of current applications in educational technologies contributes to individualization, digital and physical interaction, increasing learning performance, gaining practical experience, gamification, motivating, encouraging, improving the learning experience and increasing its effectiveness, concretization, networking, feedback, providing effective and original content, creating alternative learning processes, data-based guidance and prediction. Artificial intelligence plays an important role in making education more efficient by monitoring students through big data, making performance evaluations and providing support to teachers (Arslan, 2020). In particular, deep learning techniques provide great convenience in tailoring educational materials specific to the student (Akbulut & Adem, 2023). Gürlek et al. (2023) found in their research that it helps teachers to follow student progress, evaluate their performance and improve their teaching processes. Altun (2024) found that artificial intelligence can make significant contributions to pedagogical processes such as personalized and differentiated learning, language learning, critical thinking skills, active learning, adaptive learning, creating collaborative learning environments, creative learning, problem-based learning, project-based learning, self-directed learning, self-paced learning, self-regulated learning and simulation-based learning. Research in Turkey on the use of artificial intelligence-based systems in education has focused on university students and 178 Artificial intelligence in preschool education: Personalized assessment and evaluation with artificial intelligence USBED, 7(12), 2025, İlkbahar / Spring intelligent tutoring systems. Intelligent Tutoring Systems (ITS) are systems that support learners throughout the learning process by utilizing artificial intelligence technology. These systems guide the progress of learners with instant feedback, similar to the behavior of a tutor (Tuna, 2015). It is stated that there has been a great increase in the number of such studies in recent years and that these systems can be used at various levels from pre-school to higher education (Akdeniz & Özdinç, 2021). The scarcity of studies on the applications and effects of artificial intelligence-supported systems in education in the Turkish literature indicates that more research is needed in this field (Arslan, 2020). This study aims to improve the quality of education by introducing an innovative artificial intelligence-supported approach to the preschool education system in Turkey. In the light of these studies, the development of artificial intelligence-based educational applications in Turkey is gaining momentum. However, it is seen that there is still a limited number of studies examining the effects of artificial intelligence-supported systems in education in the Turkish literature. Increasing research in this field will contribute to the more widespread and effective use of artificial intelligence in education (Akdeniz & Özdinç, 2021). Artificial Intelligence Supported Measurement and Evaluation Model Components Natural Language Processing (NLP) Natural Language Processing (NLP) is used to assess students' language development and communication skills. NLP analyzes children's spoken and written expressions to identify skills in grammar, vocabulary and sentence structure. The use of NLP in preschool education is also increasing. For example, NLP-based practices such as interactive storytelling and educational games support children's language skills at an early age. Speech recognition technologies enable teachers to monitor children's pronunciation and speech development and help early detection of possible language delays. The use of NLP in education allows us to gain a deeper understanding of students' language skills (Oflazer, 2020; Artiwise, 2021). Image Recognition Image recognition technologies are used to monitor and assess children's physical activity and social interactions. Social cues such as facial expressions and body language can be analyzed through these technologies. Although there is a limited number of studies on the use of these 179 Gülçin GÜVEN & Seher YUMUGAN USBED, 7(12), 2025, İlkbahar / Spring technologies in preschool education in Turkey, research from around the world provides important information. In preschool education, image recognition technologies can be used to monitor children's social interactions and physical activities during group activities in detail. For example, children's motor skills can be assessed by recording their movements during play and analyzing this data (Edwards, 2018). These applications help individualize the educational process by providing teachers with important feedback on children's physical and social development. In addition, technologies based on facial recognition and emotion analysis can be used to identify children's emotional states and thus enable early intervention. These technologies contribute to the educational process by monitoring students' social interactions within the group. For example, teachers can use this data to tailor classroom activities and teaching strategies to children's needs. Machine Learning Algorithms Machine learning algorithms analyze student data to identify individual learning styles and development rates. These algorithms optimize learning processes by providing personalized feedback to each student. Several machine learning projects developed in Turkey offer individualized feedback by monitoring students' performance. For example, by analyzing children's projects and activity results, adapted learning paths can be suggested for each child (Arslan, 2020). The use of machine learning algorithms in preschool education is becoming increasingly important. With these algorithms, children's play activities and interactions can be analyzed to create educational programs tailored to their individual learning needs. For example, the data obtained from the educational games played by children on tablets or computers can be analyzed with machine learning algorithms to provide content tailored to each child's interests and learning speed (Papadakis, Kalogiannakis, & Zaranis, 2018). These applications enable children to have individualized learning experiences at an early age and increase their motivation to learn (Neumann, 2018). Furthermore, by using machine learning-based tools, teachers can better understand children's learning progress and needs and adapt their curricula accordingly (Edwards, 2018). In this way, educational processes can be optimized to contribute to the individual development of each child. 180 Artificial intelligence in preschool education: Personalized assessment and evaluation with artificial intelligence USBED, 7(12), 2025, İlkbahar / Spring Adaptive Learning Platforms Adaptive learning platforms support students' individual learning journeys and personalize educational materials. They enable more effective learning by providing content customized to students' needs. In the Turkish education landscape, projects for the automatic adaptation of personalized learning materials are becoming widespread. In preschool education, these platforms support children's learning processes by providing activities that match their interests and developmental levels (Edwards, 2018). For example, interactive educational apps and game-based learning platforms provide adaptive content to develop children's cognitive and social skills (Papadakis, Kalogiannakis, & Zaranis, 2018). Through these platforms, teachers can track each child's learning progress and offer additional support where needed (Neumann & Neumann, 2014). These applications allow children to have individualized learning experiences at an early age and increase their motivation to learn (Zuckerman, 2016). These platforms aim to manage students' educational processes based on data. Model Schematic Figure 1. Model Functioning Diagram 181 Gülçin GÜVEN & Seher YUMUGAN USBED, 7(12), 2025, İlkbahar / Spring The implementation process of the model includes data collection, analysis, personalized feedback and the use of adaptive learning platforms by integrating an AI-supported approach in education. This process is designed to monitor and support students' language development, social interactions, physical activities and overall academic performance. To measure the impact of the model in education, various quantitative and qualitative methods are used to analyze student performances, teacher feedback and changes in learning processes. This measurement process allows for close monitoring of both learning outcomes and individual student progress. Below we detail how these impacts are measured: Student Performance Data In preschool education, children's developmental processes are assessed through observation in different areas such as language skills, motor skills, social interactions and cognitive development. These assessments are based on observation data obtained from various sources such as participation in classroom activities, play-based learning activities, art and drama activities, physical activities and social interactions. By automatically analyzing this performance data, artificial intelligence-supported systems monitor children's developmental levels and help understand their individual learning processes. In particular, machine learning algorithms analyze data to determine the individual developmental rates and learning paths of each child. By comparing this data with children's developmental goals, the effectiveness of educational applications can be evaluated and personalized learning experiences can be created according to needs. Effectiveness of Personalized Feedback The impact of the personalized feedback provided by the model on children's performance is examined. To measure the impact of the feedback, children's post-feedback performance improvements are tracked. For example, assessments on vocabulary, grammar use and communication skills of a child who receives feedback in the area of language development are used to measure the impact of the model. In this process, the contribution of feedback to children's development in specific areas is analyzed by comparing before and after performances. Teacher Feedback Teachers' experiences and feedback on the implementation of the model is an important source of data. Teachers evaluate changes in children's performance, the integration of the model into the activity process and how it affects their educational process. Feedback from teachers 182 Artificial intelligence in preschool education: Personalized assessment and evaluation with artificial intelligence USBED, 7(12), 2025, İlkbahar / Spring through interviews and questionnaires is used to assess the overall impact of the model in education. This feedback provides information about the effectiveness of the model during teachers' classroom practice and indicates areas for improvement. Student Progress Monitoring Children's individual progress is regularly monitored through adaptive learning platforms. The speed at which children reach the set goals and their development processes are recorded by the platform's automatic monitoring systems. Changes in children's academic performance following adaptations to their learning processes are systematically evaluated. In particular, the differences between children's previous performances and their new performances are used to show whether the model is effective or not. Quality of Learning Experiences How children's learning experiences improve is another indicator of the impact of the model. Through questionnaires and interviews with children, it is questioned how personalized learning materials and feedback affect their learning experiences. The motivation of children to participate in activities, how satisfied they are with the learning process and how they evaluate their individual achievements are important data collected in this process. An increase in the quality of learning experiences is an important indicator of successful implementation of the model. Long-term Impacts In order to measure the long-term effects of the model in education, children need to be observed for a certain period of time. In this context, the long-term results of the model are analyzed by comparing children's periodic academic achievements and general development levels. Especially in children for whom the model is applied continuously, the sustainability of individual development and progress levels are examined. In this way, it is revealed whether the model is effective not only in the short term but also in the long term. Artificial Intelligence Tools Used in the Model The specific components in our model - natural language processing (NLP), image recognition technologies, machine learning algorithms and adaptive learning platforms - can be described as follows: 183 Gülçin GÜVEN & Seher YUMUGAN USBED, 7(12), 2025, İlkbahar / Spring Zemberek NLP (Natural Language Processing Tool) In Turkey, projects such as Zemberek have developed natural language processing systems that work on the rich morphological structure of Turkish. These systems can automatically monitor children's language development and provide developmental feedback. Zemberek can be used to analyze children's language development in our model as a natural language processing (NLP) technology specific to the Turkish language. This technology plays a major role in analyzing children's written and spoken language use to examine speech errors, vocabulary, and sentence structures (Oflazer, 2020). Teachers can use this tool to obtain data on children's language performance, especially for individualized assessment of children's writing and speaking skills. In the light of this data, targeted feedback can be provided to children who are lagging behind in language development. Sample Usage: Zemberek analyzes children's writing their names, line work, drawing pictures and tracking their linear development, and automatically assesses grammar and vocabulary usage. As a result of this analysis, teachers can identify which children need more support in terms of language development and provide customized feedback to children accordingly. Google Cloud Vision (Image Recognition Technology) Google Cloud Vision (2021), which will be used in the model, is an important tool for monitoring children's physical activity and social interactions. This technology provides teachers with data on children's motor skills and social behavior by analyzing behaviors and interactions in classroom activities (Nguyen et al., 2020). Image recognition provides detailed information about how children move and interact with other children, especially in play-based activities, allowing teachers to track their social skill development. Example Usage: Classroom activities are video recorded and analyzed with Google Cloud Vision. The tool evaluates children's movements, facial expressions and group interactions, providing teachers with data on their level of social interaction. With this data, teachers can identify which children need social skills development. Machine Learning Algorithms In our model, machine learning algorithms will be used to analyze data collected throughout children's learning process to determine their individual learning speed and learning style. Machine learning enables personalized feedback and recommendations based on learning performance (Luckin et al., 2016). By identifying children's individual learning needs, these 184 Artificial intelligence in preschool education: Personalized assessment and evaluation with artificial intelligence USBED, 7(12), 2025, İlkbahar / Spring algorithms personalize teaching processes and provide children with more effective learning experiences. Example Usage: Children's performance data (participation data, project work) is analyzed with machine learning algorithms. As a result of this analysis, learning materials and activities tailored for each child are presented and feedback is generated according to the development speed of the children. Knewton (Adaptive Learning Platform) Knewton (2021) will be used as the adaptive learning component of our model, delivering learning materials based on children's individual performance. By analyzing children's learning pathways, this platform provides the opportunity to deliver personalized educational content based on the needs of each child (Holmes et al., 2019). Thanks to the adaptive learning platform, educational materials are continuously updated in accordance with children's individual learning speeds and personalized learning paths are recommended. Example Use: The Knewton platform automatically provides personalized learning paths based on children's performance in activities. Children's development processes are regularly monitored and content is updated according to the data obtained during this process. This way, each child's learning experience becomes individualized. Faster R-CNN (Deep Learning Based Object Recognition Technology) Faster R-CNN can be used to monitor children's interactions with objects and motor skills in classroom activities. This deep learning-based technology allows assessing children's motor skill development by analyzing their interactions with objects and physical activities (Yılmaz & Aydın, 2020). Teachers can use this data to identify which motor skills children need improvement in. Example Use: Faster R-CNN analyzes video recordings of children interacting with specific objects during classroom activities. The tool assesses children's motor skills and provides teachers with data on which areas they need support. The artificial intelligence tools that will be used in our model have a wide potential to provide personalized learning experiences in education to meet the individual needs of children. Tools such as Zemberek, Google Cloud Vision, Knewton and Faster R-CNN offer powerful support to teachers in monitoring and assessing language development, social interaction, motor skills 185 Gülçin GÜVEN & Seher YUMUGAN USBED, 7(12), 2025, İlkbahar / Spring and individual learning processes. Effective use of these technologies will make education and training processes more efficient and effective. DISCUSSION, CONCLUSION AND RECOMMENDATIONS The research findings show that the personalized learning experiences offered by AI in education and its potential to optimize student performance are in line with the existing literature. Chen et al. (2020) found that AI-supported systems facilitate data-driven decisionmaking in education and improve student performance. Similarly, Holmes et al. (2019) emphasize that artificial intelligence applications are effective in personalizing educational processes and increasing student achievement. Zawacki-Richter et al. (2019) stated that artificial intelligence has changed the roles of teachers in education and enabled teachers to make more strategic interventions with data-driven learning management systems. This model design confirms that teachers have the opportunity to monitor children's developmental processes more closely and provide instant feedback through AI-supported models. The model's ability to provide personalized feedback to meet children's individual learning needs can make teaching processes more effective and efficient. Adaptive learning platforms can update educational materials by continuously monitoring student performance and thus optimize students' learning experiences (Luckin et al., 2016). AI-supported assessment and evaluation models create a more objective, consistent and reliable assessment environment by significantly limiting human errors in assessment processes. The standardized and systematic approach of artificial intelligence algorithms prevents subjective judgments and inconsistencies, allowing for a fair assessment of each student's performance. In this way, students' true potential and abilities are accurately reflected, helping them to engage more effectively in the learning process. Furthermore, AI's capacity for fast and accurate data processing allows even largescale assessments to be carried out in a short time and without errors. This allows teachers and educational administrators to save more time and energy, thus allowing them to focus more on educational activities. The ability of artificial intelligence-supported systems to minimize prejudices contributes to making evaluations on a more fair and egalitarian basis. This ensures equality of opportunity in education and allows each student to be evaluated in line with their individual characteristics. Thus, the negative effects of various socio-economic and cultural differences within the education system can be reduced. These developments pave the way for an increase in the quality of education in general and a more effective learning environment. It becomes possible for students to increase their motivation, develop positive attitudes towards learning and become an active part of their own learning processes. Educational institutions can