Опубликована: Апрель 25, 2025
Язык: Английский
Learning and Individual Differences, Год журнала: 2023, Номер 103, С. 102274 - 102274
Опубликована: Март 9, 2023
Язык: Английский
Процитировано
2454
Sustainability, Год журнала: 2024, Номер 16(7), С. 3034 - 3034
Опубликована: Апрель 5, 2024
The introduction of accessible generative artificial intelligence opens promising opportunities for the implementation personalized learning methods in any educational environment. Personalized has been conceptualized a long time, but it only recently become realistic and truly achievable. In this paper, we propose an affordable sustainable approach toward personalizing materials as part complete process. We have created tool within pre-existing management system at software engineering college that automatically generates based on outcomes provided by professor particular class. were composed three distinct styles, initial one being traditional style other two variations adopting pop-culture influence, namely Batman Wednesday Addams. Each lesson, besides delivered different formats, contained generated multiple-choice questions students could use to check their progress. This paper contains instructions developing such with help large language models using OpenAI’s API analysis preliminary experiment its usage performed 20 studying European university. Participation study was optional voluntary basis. student’s quantified, questionnaires conducted: immediately after subject completion another 6 months later assess both immediate long-term effects, perceptions, preferences. results indicate found multiple variants really engaging. While predominantly utilizing variant materials, they inspiring, would recommend students, like see more classes. most popular feature quiz-style tests used understanding. Preliminary evidence suggests various versions leads increase students’ especially who not mastered topic otherwise. study’s small sample size restricts ability generalize findings, provide useful early insights lay groundwork future research AI-supported strategies.
Язык: Английский
Процитировано
78
British Journal of Educational Technology, Год журнала: 2023, Номер 54(5), С. 1222 - 1245
Опубликована: Май 17, 2023
Advancements in artificial intelligence are rapidly increasing. The new‐generation large language models, such as ChatGPT and GPT‐4, bear the potential to transform educational approaches, peer‐feedback. To investigate peer‐feedback at intersection of natural processing (NLP) research, this paper suggests a cross‐disciplinary framework that aims facilitate development NLP‐based adaptive measures for supporting processes digital learning environments. conceptualize process, we introduce process model, which describes learners' activities textual products. Further, terminological procedural scheme facilitates systematically deriving foster how NLP may enhance adaptivity support. Building on prior research education NLP, apply all learner model exemplify range support measures. We also discuss current challenges suggest directions future effectiveness other dimensions our suggested framework, collaborations can innovate Practitioner notes What is already known about topic There considerable science processes. Natural analysis students' data. lack systematic orientation regarding techniques be applied data effectively process. adds A comprehensive overview relevant products designing An application results exemplifying use cases employed each activity during Implications practice and/or policy boost their scenarios, instructors instructional designers should identify leverage points, corresponding measures, adaptation targets automation goals based theory empirical findings. Management IT departments higher institutions strive provide tools modern models integrate them into respective management systems; those help translating requested by prediction targets, take input allow evaluating predictions.
Язык: Английский
Процитировано
70
Computers and Education Artificial Intelligence, Год журнала: 2024, Номер 6, С. 100209 - 100209
Опубликована: Янв. 25, 2024
The potential application of generative artificial intelligence (AI) in schools and universities poses great challenges, especially for the assessment students' texts. Previous research has shown that people generally have difficulty distinguishing AI-generated from human-written texts; however, ability teachers to identify an text among student essays not yet been investigated. Here we show two experimental studies novice (N = 89) experienced 200) could texts generated by ChatGPT student-written However, there are some indications more made differentiated accurate judgments. Furthermore, both groups were overconfident their Effects real assumed source on quality heterogeneous. Our findings demonstrate with relatively little prompting, current AI can generate detectable teachers, which a challenge grading essays. study provides empirical evidence debate regarding exam strategies light latest technological developments.
Язык: Английский
Процитировано
29
Telematics and Informatics Reports, Год журнала: 2024, Номер 14, С. 100127 - 100127
Опубликована: Март 7, 2024
This article presents an analysis of artificial intelligence (AI) in information systems and innovation-related journals to determine the current issues stock knowledge AI literature, research methodology, frameworks, level conceptual approaches identify gaps for future investigations. A total 85 peer-reviewed articles from 2020 2023 were used analysis. The findings show that extant literature is skewed towards prevalence technological highlights relatively lower focus on other themes, such as contextual co-creation issues, conceptualisation, application domains. While there have been increasing with intelligence, three identified areas security concern are data security, model network security. Furthermore, review found contemporary AI, which continually drives boundaries computational capabilities tackle increasingly intricate decision-making challenges, distinguishes itself earlier iterations two primary aspects significantly affect organisational learning dealing AI's potential: autonomy learnability. study contributes by providing insights into approaches, framework help
Язык: Английский
Процитировано
26
Опубликована: Май 11, 2024
Mathematical language is a cornerstone of child's mathematical development, and children can effectively acquire this through storytelling with knowledgeable engaging partner. In study, we leverage the recent advances in large models to conduct free-form, creative conversations children. Consequently, developed Mathemyths, joint agent that takes turns co-creating stories while integrating terms into evolving narrative. This paper details our development process, illustrating how prompt-engineering optimize LLMs for educational contexts. Through user study involving 35 aged 4-8 years, results suggest when interacted their learning was comparable those who co-created human However, observed differences engaged co-creation partners different natures. Overall, believe LLM applications, like offer unique conversational experience pertaining focused objectives.
Язык: Английский
Процитировано
23
Frontiers in Education, Год журнала: 2023, Номер 8
Опубликована: Дек. 5, 2023
This study reports the Intraclass Correlation Coefficients of feedback ratings produced by OpenAI's GPT-4, a large language model (LLM), across various iterations, time frames, and stylistic variations. The was used to rate responses tasks related macroeconomics in higher education (HE), based on their content style. Statistical analysis performed determine absolute agreement consistency all correlation between terms findings revealed high interrater reliability, with ICC scores ranging from 0.94 0.99 for different periods, indicating that GPT-4 is capable producing consistent ratings. prompt this also presented explained.
Язык: Английский
Процитировано
25
Computers in Human Behavior, Год журнала: 2024, Номер 158, С. 108305 - 108305
Опубликована: Май 16, 2024
This article provides a comprehensive review of current practices and methodologies within the field learning analytics, structured around dedicated closed-loop framework. framework effectively integrates various aspects analytics into cohesive framework, emphasizing interplay between data collection, processing analysis, as well adaptivity personalization, all connected by learners involved underpinned educational psychological theory. In reviewing each step closed loop, delves advancements in exploring how technological progress has expanded collection methods, particularly focusing on potential multimodal acquisition theory can inform this step. The analysis is thoroughly reviewed, highlighting range methods including machine AI, discussing critical balance prediction accuracy interpretability. personalization examines state research, underscoring significant gaps necessity for theory-informed, personalized interventions. Overall, underscores importance interdisciplinarity advocating integration insights from fields to address challenges such ethical usage creation quality experiences. aim guide future research practice promoting development effective, learner-centric environments driven balancing data-driven theoretical understanding.
Язык: Английский
Процитировано
13
Frontiers in Education, Год журнала: 2024, Номер 9
Опубликована: Июль 16, 2024
Advancements in the generative AI field have enabled development of powerful educational avatars. These avatars embody a human and can, for instance, listen to users’ spoken input, generate an answer utilizing large-language model, reply by speaking with synthetic voice. A theoretical introduction summarizes essential steps developing AI-based explains how they differ from previously available technologies. Moreover, we introduce GPTAvatar , open-source, state-of-the-art avatar. We then discuss benefits using avatars, which include, among other things, individualized contextualized instruction. Afterward, highlight challenges Major problems concern incorrect inaccurate information provided, as well insufficient data protection. In discussion, provide outlook addressing advances content technology identifying three crucial open questions research practice.
Язык: Английский
Процитировано
12
Journal of Educational Computing Research, Год журнала: 2025, Номер unknown
Опубликована: Янв. 4, 2025
Cognitive ability is closely associated with the acquisition of programming skills, and enhancing learners’ cognitive a crucial factor in improving efficacy education. Adaptive feedback strategies can provide learners personalized support based on their learning context, which helps to stimulate interest improve outcomes. Nevertheless, it remains unclear whether adaptive enhance learners. This study applies an introductory course by designing quasi-experiment analyze reveal effects program from multiple modalities, including physiological, psychological, behavioral data. Sixty-five first-year university students were randomly assigned either experimental or control groups. The group received during process non-differential feedback. findings demonstrated that significantly enhances superior performance tests, examinations, self-efficacy. Furthermore, was found markedly processing learners, as evidenced amplitude latency P300 component EEG signals, key-press reaction times accuracy rates.
Язык: Английский
Процитировано
2