Nowadays, generative AI is gradually exerting a greater influence in education. Whether it provides risks or opportunities is up to debate, and it is my conviction that it does more harm than good, especially considering the social atmosphere in Hong Kong. This essay will address the risks of generative AI, solutions that institutions in Hong Kong have come up with and their shortcomings, and methods I believe to be effective in solving AI plagiarism.
One may say that the rise of generative AI is akin to the rise of calculators; many once believed they would deteriorate mental calculation skills, yet they ended up making math more accessible, making difficult problems much easier to solve, hence accelerating mathematical breakthroughs. In the same way, one may believe that generative AI, despite worrying for now, will end up being a great tool for education. However, one must realise that a calculator cannot think of a way to solve a problem for you—it can only give you a numerical answer to the steps you have provided. You have to meticulously think about which side is perpendicular to which in a tetrahedron and find out the hypotenuse of a triangle before applying sine and cosine formulas. It is a tool that speeds up the process yet fails to complete the process on its own. In contrast, for generative AI, all students have to do is input their essay topics or schoolwork questions in a chatbot, and it would give you a full-length answer. No critical thinking is involved in this process. As professor Chiu Wing-kai from the Education University of Hong Kong says, if students submit the answers provided by generative AI tools without evaluating the validity of the answers and without any editing input from themselves, learning will become meaningless. In essence, generative AI fails to be merely a tool, like a fork to pick up a piece of food, it actively replaces the action itself by being the hand that feeds the food into one’s mouth.
Perhaps to some, the answers can be used as a guide or a reference on how to improve. Yet, the self-discipline of most youngsters is doubted. According to a global study by Best Colleges, 51% of college students believe that using AI tools like ChatGPT to complete assignments and exams is cheating, yet the same survey found that 21.5% of college students acknowledge turning to AI tools to work on assignments or exams. Among those students who used AI tools to complete assignments or exams, 17% used AI to complete an assessment and turn it in with no edits. With Hong Kong’s achievement-focused environment, it is believed that these statistics would be even more amplified in the region. The statistics, ultimately, beg the question—do most students fundamentally dislike schoolwork? And if so, why? Certainly, the ultimate goal of education is supposed to broaden one’s worldview and teach critical thinking or presentation skills, but how can the latter be achieved if one is copying an answer directly from head to toe? Say, the restrictions imposed in universities and secondary schools stop students from doing that, that still doesn’t mean that this isn’t a warning sign of the academic burden students face. It suggests that students would rather submit a model answer than to submit something that may have flaws but is truer to what the student believes. In terms of writing, there is a fundamental fear of not only failure but also uniqueness. It also exemplifies the common practice of focusing on simply getting a good score rather than learning from mistakes.
Obviously, the major concern of AI use in education is whether academic integrity can be maintained. Foreseeing the unstoppable trend, schools have taken various measures. For instance, the Chinese University of Hong Kong (CUHK) has implemented a guide of Use of Artificial Intelligence Tools in Teaching, Learning and Assessments. The 2025 version of the guide indicates that each course may adopt a different approach, such as prohibiting all use of AI tools, using AI tools only with prior permission, or using AI tools with explicit acknowledgment. Students are required to submit their completed work to CUHK’s generative AI detector VeriGuide. Violations of said regulations will be handled in accordance with the CUHK’s Procedures for Handling Cases of Academic Dishonesty, including penalties such as a failure grade for the course or reviewable/permanent demerits. However, VeriGuide has an accuracy rate of only about 80%. While false positives may be easily solved by students providing proof that they completed the assignment themselves, cheating students can simply use other AI software to rewrite the work in a way that avoids generative AI detection, such as Bypass AI, or use AI detectors to check whether their essay evades detection and edit them manually until no AI is detected, leading to a number of false negatives. It is questioned whether the reliability of detectors can be trusted and the punishments can be implemented fairly or act as an effective deterrent when there are many ways to avoid detection. These methods will only become more rampant as time goes on and technology becomes more advanced. It is also a problem if the technology of detectors cannot catch up with that of rewriting tools that avoid detection.
Some secondary schools aim to include AI in classes, and use it as teaching opportunities, such as gathering writing materials, telling students how the information that generative AI gives may not be accurate, teaching them to fact check answers, and not to abuse AI via plagiarism. The Education Bureau has also encouraged implementing AI in education, such as in the sharing session on Use of Artificial Intelligence for Enhancing Teaching Effectiveness in Science in 2025, and also launching the AI for Science Education Funding Programme. In my opinion, repackaging the role of AI as a helper or introducing AI ethics education, though a great soft strategy, is not going to change the fact that the option of cheating with AI is still readily accessible. If the academic competitiveness of Hong Kong’s society is not changed, this problem will never be fixed. The rise of AI makes us question—what is the role of education in the first place? Is it to simply learn facts or to learn how to write a good essay? And if the things we learn in school can be easily answered with AI, does that mean that we are not learning enough in schools? A simple fix would be to increase the focus of developing teenagers’ soft skills in school, such as communication skills, leadership skills, and problem-solving skills. If less time is spent on learning hard facts and more time is spent on encouraging holistic development, this may also alleviate students’ academic burden under the difficult HKDSE curriculum and reduce the incentive to simply copy an answer to get a better score. The HKEAA may also narrow down the curriculum of certain subjects with lots of things to recite such as Chemistry. Although these methods cannot solve the problem entirely, it is the first step in reshaping our social atmosphere.
Not only does the use of generative AI deteriorate critical thinking, but it also has a devastating environmental impact. A ChatGPT query consumes about five times more electricity than a simple web search. Scientists have estimated that by 2026, the electricity consumption of data centers (where answers to prompts are calculated in) is expected to approach 1,050 terawatt-hours, which would mean data centers consume more electricity than the entirety of Japan in 2024. Burning of fossil fuels used to generate electricity for data centers leads to increased carbon dioxide emissions and intensifies global warming. However, most are unaware of this environmental impact, and are not as mindful of their usage of AI, using multiple prompts to get to the point or wasting electricity to say “thank you” to a soulless machine. Schools should teach students to be aware of this and allow them to critically assess whether using generative AI is necessary at the expense of exacerbating climate change.
In sum, while AI is able to make education more accessible, it is just as capable of making cheating more accessible. It is doubted that the restrictions currently imposed can help protect academic integrity. Nevertheless, it would be harsh to say that AI should have no role in education, and it would not be feasible to ban all AI use. But before its environmental impact is alleviated and before more concrete regulations take place in secondary schools, students should monitor their AI usage and not be over–reliant. Moreover, the copying of AI answers itself leads us to question whether the school curriculum is adequate or even at the slightest bit enjoyable. The education system, especially the HKDSE curriculum, should see a fundamental shift towards promoting correct moral values and personal development instead of fostering an exam-focused generation of students.
＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿＿
Ng Wing Nam
News Commentary Competition – The Champion of Senior Form
Munsang College