International Journal of Evaluation and Research in Education (IJERE) Vol. 14, No. 3, June 2025, pp. 1761~1769 ISSN: 2252-8822, DOI: 10.11591/ijere.v14i3.33138  1761 Journal homepage: http://ijere.iaescore.com Perceptions of the generative AI-enabled cognitive offload instruction in English writing Hui Hong, Poonsri Vate-U-Lan, Chantana Viriyavejakul School of Industrial Education and Technology, King Mongkut’s Institute of Technology Ladkrabang, Bangkok, Thailand Article Info ABSTRACT Article history: Received Oct 5, 2024 Revised Feb 2, 2025 Accepted Feb 20, 2025 This study examines the students’ perceptions of the generative artificial intelligence (AI)-enabled cognitive offload instruction and its effectiveness in improving their critical thinking skills in writing English essays. This qualitative research collects data from 120 students through focus group discussions and is analyzed by Word Clouds to generate a visual representation of the word frequencies. The findings reveal that generative AI-enabled cognitive offload instruction had: i) an impact on critical thinking and writing skills; ii) effective features of Skywork, ability to generate relevant prompts and provide constructive feedback; iii) use of Skywork in developing stronger arguments; iv) promoting critical examination of different perspectives; v) interactive nature and motivation; vi) enhanced analytical skills; vii) impact on essay structuring and organization; viii) feedback and revision process; and ix) transferability of critical thinking skills. This study concludes that the highest frequency was Skywork, ability, writing, feedback, evidence, skills, thinking, arguments, essays, and peers. Students recommend in-depth explanations for complex topics, advanced tutorials, regular updates, collaboration features, advanced modules, and personalized learning paces to enhance Skyworks’s integration into instruction. Keywords: China Cognitive offload instruction Generative AI-enabled Skywork Student’s perceptions This is an open access article under the CC BY-SA license. Corresponding Author: Poonsri Vate-U-Lan School of Industrial Education and Technology, King Mongkut’s Institute of Technology Ladkrabang Chalong Krung Alley, Lat Krabang, 10520 Bangkok, Thailand Email: poonsri.va@kmitl.ac.th 1. INTRODUCTION The current situation surrounding the development of the generative artificial intelligence (AI)-enabled cognitive offload instruction approach to enhance critical thinking is characterized by a range of regulations and plan strategies published by the United Nations. For instance, the United Nations sustainable development goals (SDGs), precisely goal 4: quality education, aims to ensure inclusive and equitable quality education and promote lifelong learning opportunities for all [1]. Besides, United Nations Educational, Scientific and Cultural Organization (UNESCO) [2] is committed to supporting member states in harnessing the potential of AI to achieve the education 2030 agenda. The generative AI named “Skywork” is China’s first dual hundred-billion-scale language model, sharing similar characteristics to ChatGPT’s. It is a generative AI product developed by Kunglun Wanwei Technology [3], capable of meeting various needs such as writing essays, knowledge querying, code programming, logical reasoning, and mathematical calculations [4]. Skywork has powerful natural language processing (NLP) and intelligent interaction capabilities. It can be applied in various scenarios, such as intelligent question and answer or Q&A, chat interaction, and text generation. It has a rich knowledge base covering science, technology, culture, art, and history [5].