Recent debates over the use of artificial intelligence (AI) by students at Nanyang Technological University (NTU) have rekindled debates about what constitutes fair use of AI and Generated AI (Genai) tools in an academic setting.
To better understand this issue, we need to look at the big picture of how AI and Genai are revolutionizing the education sector in Singapore and elsewhere.
For teachers, this can meet changing classroom technical needs, including AI-specific training and technical competency training.
AI classes have already been introduced to elementary and junior high school students as part of the National AI Strategy 2.0 to expose students to Genai.
It is standard for AI to be used both in the classroom and outside the classroom, and sometimes even encourages it to improve efficiency.
But these changes require important caution – to what extent can AI be used before it is seen as overload, destructive, or even fraud?
This is the central issue behind the NTU controversy.
I accept AI, but in reality it's not
Many higher education institutions, including NTU, have allowed AI tools to be used for challenges since 2023. However, students are expected to adhere to certain guidelines when using AI in their work.
According to a statement released by NTU on June 24th, students are to declare their use of AI and how it is being used.
The university also said that some professors may ban the use of genai for certain missions. This is to “assess students' research skills and their originality and independent thinking.”
Similarly, the National University of Singapore (NUS) has detailed guidelines for the use of AI for both instructors and students.
What are the limits really?
Like some university modules Health Disease outbreak and politics Classes by Professor Sabrina Luk in the case of NTU academic fraud controversy have zero AI policy.
In this example, three students received a zero mark for class assignments after it was found to have used genai in their assignment.
Since the controversy unfolded, one student has managed to reevaluate his grades, while the other two have not been so fortunate.
Media reports suggest that students who have successfully reviewed their grades were flagged because they used a quote sorter with an expired article link and because they had three quote errors.
Citation Sorters, or Alphabetisers, are commonly used in academia, and NTU also advocates using salters such as Zotero in interdisciplinary joint core (ICC) modules, or in broad compulsory classes in schools.
These salters typically do not use Genai for their processing, but instead use programming languages such as Python.
Its use by students was deemed not to be considered genai when it was ultimately reviewed by the Academic Panel.
Additionally, schools say that when using AI, they need to “ensure de facto accuracy and properly cite all sources,” but inaccuracy is not uncommon when dealing with large quantities of quotations due to human error.
Therefore, the presence of inaccuracies in citations does not seem to be a good measure of the authenticity of the work.
Certainly, it can be argued that the rules were clearly set and that they were properly punished. AI is not permitted to be used at all. It was mistakenly detected, and the incident and subsequent damages were simply an unfortunate accident.
However, false positives of AI use in academia are not unprecedented. Students who are good at writing have been accused of using AI simply for punctuation such as short sentence structures and EM dashes.
The professor uses a variety of methods to check AI use, including AI itself, to detect AI use, but this incident raises questions about the truthfulness of the AI checker and how inappropriate AI use can be detected in the education sector.
So where do we draw the line in what is considered “acceptable” AI use? Should students avoid writing in ways that may unintentionally make them sin?
Click to subscribe to Govinsider Bulletin here.
Lack of standardization
In fact, the scope of what is considered acceptable AI use remains questionable at best.
In universities like NTU, various modules have different AI usage policies. Different allocations of the same module sometimes have different policies.
For example, the NTU History Module, HH3002 Modern East Asian Science and Technology and Medicine (the above course syllabus) allowed the use of AI in the generation of Ideas, but not a refinement of syntax and grammar using AI tools like Grammarly.
In fact, this module encouraged the use of AI to generate ideas and assist in the acquisition of primary and secondary resources, as long as the detailed prompts used have been submitted.
This is in stark contrast to the guidelines of classes that were caught up in university controversy. However, both modules are part of NTU's comprehensive college, arts and social sciences (Cohass), with similar emphasis on writing and research.
In particular, one student involved in the controversy said that grammar was not considered AI by professors, but is considered to be that way. HH3002.
Obviously there is a clear lack of standardization regarding AI usage policies.
We can't help but wait for greater standardization for what is considered AI.
The missing pieces
AI use cases such as language correction have become very common among students, especially in higher education, raising the overall baseline for excellent writing.
While poor people in grammar, spelling and vocabulary may have hindered some students in the past, AI has played a role in leveling the arena.
A study conducted by Harvard Business School alongside MIT and Wharton shows that AI is pushing the performance curve to the right and increasing the “average” of the overall job. More people are working better.
For the average student at the odds of the bell curve, not using AI in their work means they leave behind their peers.
It's ironic that some schools are promoting AI adoption, but at the same time it's hard to ignore punishing students for simply being involved with or viewing them as AI.
Progress
In an age where AI is no longer a future possibility, but now a reality, the idea of limiting AI, as NTU stated, makes “the student's research skills and make originality and independent thinking more accurately “originality and independent thinking.”
AI is rooted in everything, beyond the education sector.
Last November, the use of AI-generated artwork by local bank DBS was considered to cause anger among local artists and harm the artist's livelihoods.
The same problem occurred when online users began to flood the internet with images of Studio Ghibli-inspired AI-generated art.
Still, whether we accept it or not, AI is the future. Students use it whether it is permitted or not, and companies use it as long as profits are viable.
Schools need to prepare their students for the real world. This is a world that is changing very rapidly.
Like a course syllabus HH3002 The NTU module acknowledges that “artificial intelligence tools are constantly changing the way we research and produce research.”
It is ridiculous to impose standards in an era when AI was nothing more than our imagination of our present reality.
Similarly, technology is rapidly advancing, and society is now reaching its fourth industrial revolution.
In the 1980s, American teachers protested the use of calculators in classrooms. Today, calculators have become an integral part of mathematics and science education.
Although computer policies approved by the Singapore education system are actually in place, they have evolved to improve the balance of efficiency by supporting the development of mathematical thinking and problem-solving skills.
Perhaps instead of questioning whether AI should be permitted or educated, we need to reconsider the line of “acceptable” AI use to accommodate the rapid changes in AI.
Instead of selectively exacerbating AI, it is best to allow schools to embrace it and teach students how to use AI productively to complement their intelligence.
