Around February 12, 2026, the UAE Ministry of Education released a manual titled “2026 Safe and Responsible Use of Artificial Intelligence in Classrooms,” drawing hard lines around generative AI in schools: students under 13, or below Year 7, are banned outright from using generative AI tools. Gulf News’s report, updated February 14, and Times of India’s follow-up — headlined around the “25 new AI guidelines” — both trace back to the same document.
This is not another set of aspirational principles. The UAE produced classroom-level rules that can actually be enforced, and the timing is deliberate: from August 2026, every government school and MoE-curriculum private school must teach AI education from kindergarten through Grade 12, across seven core areas. Banning and mandating at once makes this the most concrete K-12 AI governance specimen of the young year (see our 2026 opening outlook). It also lands while regulators elsewhere are still arguing over definitions, which is precisely why it will get studied.
What the Rules Say
The core line is age. The ministry’s rationale is protecting younger learners, keeping that stage of education anchored in “interaction, creativity and independent skill development.” Anyone under 13 or in a grade below Year 7 cannot use generative tools, full stop, regardless of the purpose.
The rest of the manual spreads across roughly 25 prohibitions. The hardest one: no AI during formal exams or official assessments. Assignments, research, or reports produced with AI must be disclosed and teacher-approved; presenting them as personal work without that is a violation. Copying or rephrasing AI output without genuine understanding is banned too.
The Logic Behind the Red Lines
Laid out, the prohibitions fall into four buckets. Content: no generating or sharing violence, hate speech, discrimination, misinformation, psychologically harmful material, or anything conflicting with “religious, national or cultural values”; content encouraging cyberbullying, self-harm, or illegal behavior is out. Personal data: no feeding names, photos, audio, video, or contact details of students, teachers, or parents into AI systems. Safety: no deepfakes, impersonation, collecting sensitive information about others, accessing confidential school materials, or recording and transcribing classes without every party’s consent. Access: no unapproved platforms, no bypassing firewalls, no VPNs to evade restrictions, and no creating student accounts without permission.
For anyone building education products, the list doubles as a requirements document: age gating, content classification, data protection, and platform allowlists each map to a buildable feature.
The Teacher Stays in the Loop
Every permitted use shares one precondition: direct teacher oversight with clear instructional guidance. Teachers and students alike must verify accuracy, relevance, and alignment with approved curricula before use. The official framing is deliberately modest — AI remains “a supportive educational tool rather than a substitute for teachers or students’ intellectual effort.”
Penalties introduce no new law; misuse falls under accountability within existing educational regulations. That keeps enforcement inside school administration instead of building a separate AI police. It also means the manual’s practical weight depends on principals and teachers interpreting it consistently — the same population the rules task with supervising every permitted use.
From Banned to Mandatory
The same policy framework carries the opposite signal: starting August 2026, AI education becomes mandatory from KG to Grade 12 across seven core areas. In other words, the UAE will teach 4-year-olds about AI while barring them from using generative tools until 13. Teach understanding first, operation later — the sequencing itself is a position.
One level out, the rules sit under the UAE’s National AI Strategy 2031, making this the Gulf’s first batch of governance detail that reaches all the way down to the classroom.
Notes for EdTech Builders
Three practical effects. First, products entering UAE school markets need built-in age tiering and teacher-supervision workflows or they will not clear procurement. Second, the personal-data red line effectively bars sending student data to any third-party model — on-prem or tightly isolated deployment becomes a hard requirement. Third, “25 prohibitions”-style concrete rules travel across borders far more easily than abstract principles; expect neighboring markets to copy-paste, and teams with compliance architecture ready will move fastest.
Sources
- UAE sets strict new rules on generative AI in schools, banning use for under-13s — Gulf News
- UAE issues 25 new AI guidelines for schools, prohibits use for under-13 students — Times of India
- UAE Ministry of Education school AI rules — Lexis Middle East
AI-assisted summary compiled from the sources above, reviewed by a human before publishing.