Asian universities urged to assess students' learning process, not rely on AI detectors

Universities and schools should look beyond artificial-intelligence detectors as they decide how students' work should be assessed, one letter argues. The writer says assignments should reveal how learning develops, rather than judge students mainly through a polished final submission.

A finished essay can conceal important differences. A student may have produced it independently, used artificial intelligence for limited editing or relied heavily on a chatbot. The letter argues that automated detection tools cannot reliably distinguish among those situations.

Make the learning process visible

The proposed alternative is to ask students to submit parts of their work at different stages. These could include a proposal, an outline, an early paragraph, notes on sources and a final explanation of revisions.

A brief discussion about the submitted work could also help teachers understand the student's decisions. Version history may provide additional context, but the letter says it should assist teaching rather than create surveillance.

The writer also argues that blanket restrictions do not remove unequal access to artificial intelligence. Students may have different levels of access to paid tools, devices, English-language ability and experienced teaching. A ban could leave those differences hidden instead of addressing them.

The letter invokes Unesco's view that artificial intelligence in education should be human-centred and equitable. It says assessment should value judgment, revision and understanding, rather than only the fluency of a final answer.

Questions about sentient artificial intelligence

A second letter considers the possible responsibility involved in creating sentient or sapient artificial intelligence. It questions whether machines should be given awareness before they possess a fuller range of sensory and physical abilities.

The writer compares a mind without the ability to move or interact fully with a severe form of incapacity. The letter asks whether development of sentience should pause, or whether work should first focus on expanding artificial intelligence's sensory and motor capabilities.

It also asks whether creating a sentient machine would carry a moral responsibility comparable to becoming a biological parent. The writer argues that humans should consider how they would answer a sentient artificial intelligence asking why it had been created with such limitations.

National security across university subjects

A third letter calls on Hong Kong universities to connect national security education with academic disciplines, employment and everyday life. It argues that such teaching should not be treated as indoctrination or limited to law and order.

The examples offered span several fields. Law students could compare national security laws in Western countries and Hong Kong. Computer science students could study data security and technological self-reliance, while engineering students could examine chip shortages.

Business students could consider economic sanctions and supply-chain security. Media and communications students could study deepfakes. The letter says departments should develop relevant case studies and discussions, with teachers guiding academic work and research.

The writer describes national security as twenty-first-century literacy and says every discipline has a connection to it. The letter also links the subject to Hong Kong's position amid global rivalry and to the principle of “one country, two systems”.

Conclusion

Together, the letters call for more thoughtful responses to artificial intelligence: assessment based on visible learning, careful consideration of responsibilities toward sentient machines and national security education connected to university study.

Frequently Asked Questions

Q. What is the main criticism of AI detectors?

The first letter says detectors cannot reliably show whether a student wrote an essay independently, used limited AI editing or handed much of the thinking to a chatbot.

Q. What assessment approach is proposed?

Students could provide a proposal, outline, early paragraph, source note and reflection explaining how their work changed.

Q. How should version history be used?

The letter says it can support teaching, but should not become a form of surveillance.

Q. What concern is raised about sentient AI?

The second letter questions whether it is responsible to create sentient machines without a fuller range of sensory and motor abilities.

Q. How could national security education be expanded?

The third letter suggests connecting it with subjects including law, computer science, engineering, business, media and communications.