Core Principles and Universal AI Ethics
Five principles underpin every practice and policy in the guide. Six ethical issues sit alongside them, and a shelf of legal instruments behind both.
The five core principles
Inclusivity and Equityp.37
AI is a tool to close the digital divide and give fair access and benefit to all students including those with special education needs. Every student, regardless of ability, gets the same opportunity to participate actively and meaningfully.
Accountability and Transparencyp.37
Humans remain responsible for designing, training and supervising AI. Every final decision requires human verification and judgement, because AI operates on pre-set algorithms.
Data Security and Privacyp.37
Protecting student and educator personal data is the highest priority. Users must be informed of the scope and limits of personal-data use; they must always check application privacy policies and never share data carelessly.
Responsible Innovationp.38
Creativity and ethical exploration with AI are strongly encouraged — while avoiding misuse such as plagiarism, copyright infringement or manipulation.
AI Does Not Replace the Human Rolep.38
Human-in-the-loop. AI supports and eases tasks but does not replace human roles, thinking or interaction. Human relationships remain the most important element in teaching and learning.
Six ethical issues to watch
(a) Academic Honesty and Plagiarismp.38
The main issue is copying or completing work without one's own effort. AI must be used as a learning aid to produce original work reflecting academic integrity.
(b) Algorithmic Biasp.38
AI output can be biased because AI learns from existing data that may contain stereotypes. Cultivate critical thinking about AI content and ensure the applications used are not biased.
(c) Information Integrity and Digital Authenticityp.38
Beware deepfakes — highly realistic but fake multimedia made to mislead or manipulate. Uncontrolled use breaches the ethical principle of honesty and can cause defamation and reputational damage.
(d) Information Accuracyp.39
AI can be misused to spread falsehoods and can make mistakes — hallucination. Fact-checking is essential; AI output must never be accepted wholesale without verification.
(e) Data Security and Privacyp.39
Open or public AI applications risk exposing personal information. Every user is responsible for not sharing sensitive information such as full name, address or phone number.
(f) Policy and Legal Compliancep.39
All stakeholders must refer to the circulars and guidelines in force, to keep AI use lawful, prevent falsehoods and safeguard national sovereignty.
What the guide weighs up
Six benefits p.20
Automating educator tasks
AI takes over repetitive work such as marking objective quizzes and managing data, so educators can focus on creativity and human interaction.
Personalised and differentiated learning
Materials and activities tailored to each student's interest and ability, including specific support for students with special education needs.
Fast and accurate
Processes and analyses large volumes of data quickly for more accurate decisions and predictions.
Early intervention
Analytics help educators identify learning problems early and suggest interventions to prevent dropout.
Unlimited availability
Educational chatbots operate 24 hours a day — help at any time.
Accelerating research
Speeds up research: analysing data, identifying patterns, running simulations.
Five risks p.21
(i) Academic integrity
Plagiarism and cheating risk rises if students use AI to complete work without their own effort or thought.
(ii) Algorithmic bias
AI trained on biased data — gender or racial stereotypes — will repeat and amplify that bias in its output.
(iii) The digital divide
Unequal access to devices and quality internet can widen the achievement gap.
(iv) Data security and privacy
Third-party AI platforms raise serious concerns about how student personal data is collected, used and protected.
(v) Over-reliance
Excessive dependence harms critical thinking, basic skills and fundamental human relationships.
The rise of deepfakesp.21
Media-manipulation technology using deep-learning algorithms to produce highly realistic fake audio, images or video. In education it risks spreading false information and damaging individual safety and reputation.
The human role is irreplaceablep.21
However powerful AI is, the human role remains and cannot be replaced. Users are responsible for designing, training and supervising AI so it is used accurately, lawfully, ethically and safely.
The legal shelf
Fourteen instruments the policy names by title. These are what turn guidance into obligation.
Education Act 1996
Act 550Sections 134 and 135 set student conduct, discipline and the powers of the headmaster. AI-related misconduct can be treated as a breach of school rules.p.10
Education (School Discipline) Regulations
1959School discipline regulations.p.10
Personal Data Protection Act 2010
Act 709Regulates personal data in commercial transactions. Section 3 exempts the Federal and State Governments — but Ministry staff must still ensure awareness of the limits of personal data use, especially with third-party AI platforms.p.10
Cyber Security Act 2024
Act 854A comprehensive cybersecurity legal framework: protection of critical infrastructure, regulation of cybersecurity services, incident-reporting procedures and continuous threat monitoring.p.11
MOE Cyber Security Policy v2.0
June 2024Ministry-specific policy binding on all Ministry staff, vendors and third parties using Ministry ICT assets — explicitly including AI technology.p.11
Online Safety Act 2025
Act 866Obligations on online application service providers to protect children — safe design and operation of online services.p.11
Education (National Curriculum) Regulations 1997
P.U.(A)531/1997Lesson plans built with AI must stay compliant.p.11
Education (Assessment and Examination) Regulations 1997
P.U.(A)533/1997Assessment adapted with AI must stay compliant.p.12
MOE Professional Circular No. 11 of 2023
4 Oct 2023Guidelines on Values and Ethics in Social Media Use.p.13
Public Sector AI Adoption Guidelines
2025Whole-of-government guidance on adopting AI.p.13
Communications and Multimedia Act 1998
Act 588Governs online conduct.p.13
Intellectual property statutes
1983 / 1996 / 2019Patents Act 1983, Industrial Designs Act 1996 and Trademarks Act 2019 — copyright and IP in AI-generated work.p.13
Public Officers (Conduct and Discipline) Regulations
P.U.(A) 395/1993Sub-regulations 19(1) and 19(2) — the prohibition on making public statements.p.13
Government Contracts Act 1949
Act 120Any cooperation with strategic partners must comply with this and the 2024 Cabinet Secretary's Directive.p.62