Where Generative AI Fits Within and in Addition to Existing AI K12 Education Interactions: Industry and Research Perspectives
Recent developments in Generative AI have led capital market, industry, and research institutions to explore its education applications as solutions to K12 challenges. However, there is currently a gap of analytical review of these trends. This chapter attempts to review and analyze predominant Generative AI education efforts within and in addition to existing AI education frameworks that include contributions from both industry and research institutions. Our aim is to present a holistic review of AI Education key interactions, explore the opportunities that Generative AI presents, share industry experience in implementing Generative AI in AIED products and identify future work directions.
Adoption of Artificial Intelligence in Schools: Unveiling Factors Influencing Teachers’ Engagement
The slow adoption of AI-based adaptive learning platforms in schools prompts concerns about identifying and predicting factors influencing teachers' engagement, leading to the development of a reliable instrument measuring holistic adoption factors; results reveal the importance of factors such as workload, ownership, trust, support mechanisms, and ethical considerations alongside knowledge, confidence, and product quality, concluding that addressing these factors can enhance the real-world adoption and effectiveness of adaptive learning platforms in schools.
AI Applications and Implications in the UAE K12 Sector: Industry Experience from Alef Education
"This paper explored the impact of AI and globalization on education systems, particularly in the UAE, highlighting the need to prepare children for future jobs, train teachers and school leaders for emerging technology trends, and understand the tasks and skills influenced by AI within the education system, presenting insights from Alef Education's AI applications and discussing the implications of AI on human development in education.
"
Explaining transformer-based models for automatic short answer grading
Recent advancements in natural language processing have made advanced language models easily accessible, but achieving infallible Automatic Short Answer Grading (ASAG) systems remains challenging; this work proposes a framework to select the most accurate and intuitive models for ASAG by evaluating popular transformer-based models with explainability methods on the Semeval-2013 benchmark dataset.