SỬ DỤNG TRÍ TUỆ NHÂN TẠO TRONG NGHIÊN CỨU KHOA HỌC: CƠ HỘI, THÁCH THỨC VÀ VẤN ĐỀ LIÊM CHÍNH HỌC THUẬT

Các tác giả

  • Đào Trung Hiếu

DOI:

https://doi.org/10.59266/houjs.2025.667

Từ khóa:

trí tuệ nhân tạo, nghiên cứu khoa học, liêm chính học thuật, đạo văn, chuẩn mực quốc tế

Tóm tắt

Bài viết phân tích một cách hệ thống cơ hội, thách thức và các vấn đề liêm chính học thuật phát sinh từ việc ứng dụng trí tuệ nhân tạo (AI) trong nghiên cứu khoa học hiện đại. Trên cơ sở khảo sát các xu hướng công nghệ và chuẩn mực học thuật quốc tế, bài viết làm rõ những tác động đa chiều của AI tới chu trình sản xuất tri thức, đồng thời đề xuất giải pháp nhằm sử dụng AI có trách nhiệm trong nghiên cứu, bảo đảm tính trung thực và minh bạch khoa học trong bối cảnh chuyển đổi số.

Tài liệu tham khảo

[1]. Committee on Publication Ethics (COPE). (2023). Authorship and AI tools: How should editors handle “AI- generated” submissions. Retrieved August 23, 2025, from https:// publicationethics.org

[2]. Crothers, B., & Zhang, L. (2023). AI misuse in higher education. Nature Human Behaviour, 7(5), 615-618. https://doi.org/10.1038/s41562-023- 01678-2

[3]. Elsevier. (2023). Publishing ethics & policies on generative AI. Elsevier. Retrieved August 23, 2025, from https:// www.elsevier.com/about/policies-and- standards/publishing-ethics

[4]. Elsevier Research Intelligence. (2024). Artificial intelligence in research: Global trends and insights. Elsevier. Retrieved August 23, 2025, from https://www.elsevier.com/research- intelligence

[5]. European Parliament. (2024). Artificial Intelligence Act adopted - key provisions and timeline. European Union. Retrieved August 23, 2025, from https://www.europarl.europa.eu/news/ en/press-room/20240313IPR20339/ artificial-intelligence-act

[6]. EUR-Lex. (2024). Regulation (EU) 2024/1689 of the European Parliament and of the Council on artificial intelligence (AI Act). Retrieved August 23, 2025, from https://eur- lex.europa.eu/legal-content/EN/ TXT/?uri=CELEX%3A32024R1689

[7]. Floridi, L., & Chiriatti, M. (2023). GPT and the ethical challenges of AI in science. AI & Society, 38(2), 231-240. https://doi.org/10.1007/s00146-023- 01564-6

[8]. Hutson, M. (2021). Robo-writers: The rise and risks of language-generating AI. Nature, 591(7848), 22-25. https:// doi.org/10.1038/d41586-021-00530-0

[9]. OpenAI. (2023). AI Classifier is no longer available. OpenAI. Retrieved August 23, 2025, from https://openai. com/blog/new-ai-classifier

[10]. RMIT Vietnam. (2023). Complexities of generative AI policies in higher education. RMIT University Vietnam. Retrieved August 23, 2025, from https://www.rmit.edu.vn/news

[11]. Springer Nature. (2023). Advice on the use of AI tools in manuscript preparation & peer review. Springer Nature. Retrieved August 23, 2025, from https://www.springernature.com/ gp/policies/editorial-policies/ai

[12]. Stanford Institute for Human-Centered Artificial Intelligence (HAI). (2023). On the (un)reliability of AI-generated text detectors. Stanford University. Retrieved August 23, 2025, from https:// hai.stanford.edu/news/reliability-ai- generated-text-detectors

[13]. Stanford Institute for Human-Centered Artificial Intelligence (HAI). (2024). AI Index Report 2024. Stanford University. Retrieved August 23, 2025, from https://aiindex.stanford.edu

[14]. Turnitin. (2023). AI writing detection: Accuracy, false positives, and guidance for educators. Turnitin. Retrieved August 23, 2025, from https://www.turnitin.com/products/features/ai- writing-detection

[15]. United Nations Educational, Scientific and Cultural Organization (UNESCO). (2023). Guidance for generative AI in education and research. UNESCO Publishing. Retrieved August 23, 2025, from https://unesdoc.unesco.org/ ark:/48223/pf0000387242

[16]. Vietnam National University (VNU). (2024). VNU to introduce AI courses for students. VNU News. Retrieved August 23, 2025, from https://news.vnu.edu.vn

[17]. Voice of Vietnam (VOV). (2023, March 15). ChatGPT vào đại học: dùng sao cho đúng? VOV Online. Retrieved August 23, 2025, from https://vov.vn/giao-duc/chatgpt-vao-dai-hoc-dung- sao-cho-dung

[18]. VnExpress. (2023, April 2). ChatGPT bùng nổ - trường học cảnh báo đạo văn và kỹ năng số. VnExpress. Retrieved August 23, 2025, from https://vnexpress.net/chatgpt-bung-no-truong- hoc-canh-bao-dao-van-va-ky-nang- so-4593561.html

[19]. Wiley Author Services. (2023). Best practice guidelines on generative AI for authors and editors. Wiley. Retrieved August 23, 2025, from https://authorservices.wiley.com/author-resources/Journal-Authors/Prepare/ writing-and-ethics/best-practice- guidelines-on-generative-ai.html

[20]. Wired. (2024, February 21). A top science journal retracted a paper after AI concerns highlighted deeper editorial flaws. Wired. Retrieved August 23, 2025, from https://www. wired.com/story/ai-science-journal- retraction

[21]. World Association of Medical Editors (WAME). (2023). Chatbots, generative AI, and scholarly publications (recommendations). WAME. Retrieved August 23, 2025, from https://wame. org/page3.php?id=106

Phụ lục:

Phụ lục A - Ma trận “đèn giao thông” sử dụng AI trong nghiên cứu:

Màu xanh (Được phép): Các hoạt động như kiểm tra chính tả, sửa lỗi ngữ pháp, dịch thô, gợi ý dàn ý, tạo mã minh họa đều được chấp nhận, miễn là có khai báo trong Lời cảm ơn hoặc Phương pháp.

Màu vàng (Hạn chế): Các hoạt động như sinh văn bản học thuật dài, tóm tắt kết quả nghiên cứu, gợi ý giả thuyết hoặc sinh mã phân tích dữ liệu chỉ nên dùng trong phạm vi hỗ trợ. Khi sử dụng, tác giả phải kiểm chứng nguồn, lưu lại nhật ký AI và có xác nhận của đồng tác giả.

Màu đỏ (Cấm): Nghiêm cấm để AI viết thay phần kết quả và thảo luận, tạo dữ liệu giả, sinh trích dẫn không kiểm chứng, tải bản thảo hoặc phản biện kín lên công cụ AI công khai. Đồng thời, không được dùng công cụ phát hiện AI như bằng chứng duy nhất để kỷ luật (COPE, 2023; WAME, 2023; Science, 2023; Stanford

HAI, 2023; Turnitin, 2023).

Phụ lục B - Kế hoạch triển khai khung AI có trách nhiệm (6 - 12 tháng)

Tháng 1 - 2: Thành lập Ban Công nghệ & Liêm chính học thuật; rà soát quy chế hiện hành; ban hành chính sách tạm thời dựa trên ma trận “đèn giao thông”.

Tháng 3 - 4: Tập huấn giảng viên - nghiên cứu sinh; triển khai biểu mẫu khai báo AI; quy định cấm tải bản thảo hoặc phản biện kín lên công cụ AI công khai.

Tháng 5 - 6: Tích hợp công cụ hỗ trợ (kiểm lỗi, quản lý trích dẫn); xây dựng Sổ tay kiểm chứng trích dẫn; thử nghiệm tổ kiểm định chéo “human-in-the-loop”.

Tháng 7 - 12: Đánh giá - cập nhật chính sách; ban hành quy chế chính thức;

công bố báo cáo thường niên về liêm chính học thuật & AI.

Phụ lục C - Một số công cụ AI tiêu biểu dùng trong học thuật

Hỗ trợ đọc - tổng quan: Semantic Scholar (AI), scite_, Elicit.

Viết - biên tập: Grammarly, Writefull, LanguageTool, DeepL Write.

Phân tích dữ liệu/mã: Jupyter Notebook với Copilothoặc CodeWhisperer (chỉ cho mã không nhạy cảm).

Quản lý trích dẫn: Zotero, EndNote (cảnh giác chức năng “tự sinh trích dẫn”).

Kiểm tra tương đồng & AI: Turnitin (không dùng làm bằng chứng duy nhất); không khuyến nghị dựa vào “AI detector” thuần túy.

Tải xuống

Loading...