Promoting the application of purely Vietnamese AI in teaching, learning, and research.

NDO - According to experts, the use of AI that is not purely Vietnamese in teaching and research in Vietnam may have disadvantages in terms of language, culture, and content. Therefore, it is necessary to promote the development of large-scale language models from open models for practical application, especially in education.

In implementing the Youth Month plan for 2025, on March 28th, the Propaganda Department of the Youth Union of the Ho Chi Minh National Academy of Politics , in coordination with the Youth Union of the Research-Teaching Bloc, organized a scientific seminar on the topic "AI in teaching and research in Vietnam today".

The scene at the seminar. (Photo: VAN TOAN)

Attending the seminar were Associate Professor, Dr. Nguyen Duc Minh, Vice Rector of the School of Electrical and Electronics Engineering, Hanoi University of Science and Technology; Mr. Dinh Trung Son, Deputy Secretary of the Youth Union of the Ho Chi Minh National Academy of Politics; and a large number of officials, lecturers, and youth union members from the entire academy system.

The speakers at the panel discussion were two experts from AI Hay company: Dr. Nguyen Tho Chuong - Chief Technology Officer and Mr. Nguyen Hoang Hiep - Chief Operating Officer.

AI Hay is an AI-powered search and data analysis tool developed entirely by a team of Vietnamese engineers, with the goal of creating an "AI product by Vietnamese people - for Vietnamese people". After nearly four years of development, the application has over 15 million users, more than half of whom are students.

Sharing his thoughts on the potential applications of AI generation in research and teaching, Mr. Nguyen Hoang Hiep, Director of Operations at AI Hay, stated that AI generation can support the creation of lectures; generate learning content ( videos , audio, text, etc.) and homework assignments, tests; automate administrative tasks, analyze student data, etc. Simultaneously, it effectively supports summarizing and comparing documents; generating ideas and building hypotheses; analyzing and interpreting data; and assisting in writing research papers.

Mr. Nguyen Hoang Hiep, Director of AI Hay Operations, giving a speech. (Photo: VAN TOAN)

However, AI Hay representatives also noted several major issues when using foreign AI generation tools, such as data security; lack of localization, non-Vietnamese characteristics, and language barriers; the default knowledge of these models may not be suitable for Vietnam; lack of content filters specific to the culture or policies; inability to control the source of information when responding to results; platform instability, and regulations that can change at any time…

"Currently, the applications we commonly use do not allow Vietnamese people to be independent in terms of technical platforms, data, culture, and information sources. Furthermore, we cannot effectively use the Vietnamese language in advanced reasoning models," said Mr. Nguyen Hoang Hiep.

He noted that the AI ​​generation mechanism, which predicts the next token (word), is suitable for tasks such as extended summarization and the reconstruction of prior knowledge. However, caution is needed when using AI generation for tasks requiring objectivity, emotion, and factual observation.

The use of AI in non-Vietnamese contexts in research and teaching in Vietnam may present disadvantages related to language, culture, and content. Therefore, it is necessary to promote the development of large-scale language models specifically from open models for practical application, especially in education.

Associate Professor, Dr. Nguyen Duc Minh, Vice Rector of the School of Electrical and Electronics Engineering, Hanoi University of Science and Technology, shared his insights at the seminar. (Photo: VAN TOAN)

Agreeing with this viewpoint, Associate Professor Dr. Nguyen Duc Minh, Vice Rector of the School of Electrical and Electronics Engineering, Hanoi University of Science and Technology, also believes that using a non-Vietnamese AI in teaching and research is an extremely serious issue.

According to him, AI generative technology is developing at a breakneck pace and is being increasingly applied in daily life, placing Vietnam at great risk of losing its sovereignty in the cultural and social fields.

"AI trained on people's data will speak in people's voices. Especially for those working in the field of political theory, mastering the 'voice of Vietnam' is extremely important; if they don't master it, they will lose their intellectual property," Associate Professor Dr. Nguyen Duc Minh noted.

Another issue raised by this expert is the ethical implications of using AI in research, learning, and teaching. The question is whether the content generated is a product of user thought or AI thought; this distinction is crucial.

Dr. Nguyen Tho Chuong, Director of AI Technology Hay, shared this. (Photo: VAN TOAN)

Speaking about the challenges of applying AI, Dr. Nguyen Tho Chuong, Director of AI Technology at Hay, said that it is very difficult to verify the quality and origin of the content provided by AI, as well as the transparency and selection of information that aligns with the nation's development direction. On the other hand, AI often collects and processes large amounts of users' personal data, which poses a significant challenge to security and privacy.

To address these issues, the AI ​​Hay application was developed with the ability to accurately process natural language in Vietnamese, integrate real-time data from reliable sources, and provide diverse learning support tools.

The highlight of AI Hay is its system for organizing authentic information sources, tailored to Vietnamese culture, capable of producing highly reliable answers. To address the issue of "AI illusions" (AI inventing new, fictitious information), AI Hay uses filters to ensure the information is clean and trustworthy.

“Where does this information come from? Are the sources contradictory? Is this answer appropriate to the context? These are the layers of logic that AI Hay employs, hoping that when an answer is given, it will come from a reliable source and be cited from that source,” said Dr. Nguyen Tho Chuong.

Discussing the application of AI Hay in teaching and learning, Dr. Nguyen Tho Chuong stated that AI Hay makes lesson preparation and information gathering easier and more intuitive for learners, systematizing their knowledge in the form of mind maps; providing summaries of lessons, with specific sources for each piece of information; and designing lectures according to requirements…

Comrade Dinh Trung Son, Deputy Secretary of the Youth Union of the Ho Chi Minh National Academy of Politics, speaks at the seminar. (Photo: VAN TOAN)

While agreeing that AI can help suggest ideas, some delegates argued that in scientific research, what matters most is not just the idea itself, but also its novelty. AI doesn't necessarily bring novelty to scientific research; it is the users who develop the idea into something new that creates value for the research topic.

Delegates emphasized the need to master AI, viewing it only as a supporting tool and not becoming completely dependent on it. In particular, caution is needed when using AI in political theory teaching; over-reliance on AI and the resulting inaccurate information could be very dangerous for public discourse.

At the seminar, experts, speakers, and delegates exchanged and discussed how AI can be applied in lesson planning, document searching, lecture design, and delivering teaching content to learners; which AI applications can be used for brainstorming, outlining, and checking for duplication in scientific tasks; copyright and intellectual property issues related to products created using AI; and how to formulate queries to provide information that most closely matches the content users are looking for…

Source: https://nhandan.vn/thuc-day-ung-dung-ai-thuan-viet-trong-giang-day-hoc-tap-va-nghien-cuu-post868312.html