AI in Schools Needs More Than a Ban or a Free-for-All

What is the current debate about AI in schools?

The debate is moving beyond whether tools such as ChatGPT should be permitted. The more useful question is how schools can introduce AI without confusing assistance with learning. A September 21 report by The Times described criticism from educator Sir Anthony Seldon and investor Tim Bunting, who argued that the United Kingdom needs a faster and more coherent national approach.

Their concern reflects an international problem. AI is already available to teachers, students and parents, but policies, training and assessment practices are developing unevenly.

What can AI do well in education?

AI can explain a difficult concept in several ways, generate practice questions, help teachers prepare differentiated materials and support students who need translation or accessibility tools. It can also reduce administrative work such as drafting routine communications or organising lesson resources.

These benefits are strongest when AI supports a defined learning objective. A student who asks for another explanation of fractions may gain understanding. A student who submits an AI-generated answer without reading it may bypass the learning process entirely.

What are the main risks?

Incorrect information

Generative AI can produce convincing but inaccurate statements. Students may not have enough subject knowledge to recognise an error. Teachers therefore need to treat generated material as a draft requiring verification.

Academic integrity

Traditional homework can become difficult to interpret when students use AI without disclosure. Schools need rules that distinguish permitted assistance—such as brainstorming or grammar support—from prohibited substitution of the student’s work.

Privacy

Students and teachers may enter names, grades, behavioural records or health information into external tools. Personal data should not be uploaded unless the school has approved the system, assessed its terms and established a lawful and secure process.

Over-reliance

Students still need memory, reasoning, writing and problem-solving skills. If AI supplies every first draft and every answer, learners may lose opportunities to practise.

Unequal access

Paid tools, better devices and faster internet can create advantages unrelated to ability. A school AI policy should include access and fairness, not only discipline.

What should a responsible school framework contain?

How should schools change assessment?

Assessment should measure understanding rather than the ability to produce polished text at home. Teachers can ask students to explain their reasoning, submit draft stages, cite tool use or defend a project orally. In-class writing remains useful, but it should be one part of a wider evidence set.

Schools should avoid relying entirely on AI-detection software. Detection systems can produce false positives and should not be treated as conclusive proof of misconduct.

What does this mean for Philippine schools?

Philippine schools face the same core issues, with additional differences in connectivity, device access and teacher workload. Policies should be practical enough for public and private institutions with different resources.

A sensible starting point is a three-level system: allowed uses, uses requiring teacher approval and prohibited uses. For example, translation of instructions may be allowed; AI assistance on an assessed essay may require disclosure; uploading student records to a public chatbot should be prohibited.

Conclusion

AI can improve education when it strengthens teaching and gives students more ways to understand. It becomes harmful when it replaces thinking, exposes personal data or makes assessment meaningless. Schools need neither a total ban nor unrestricted use. They need a clear educational purpose, approved tools, human accountability and rules students can understand.