The UAE AI Curriculum Fits in 36 Periods. The Training Does Not.
HONG KONG, Nov 4, 2026 - (ACN Newswire) - A Grade 7 teacher in Sharjah opens her timetable for the week beginning 31 August 2026, the first week of the new academic year. She has 36 periods to fill, the allocation the Ministry of Education sets for the second and third cycles. Arabic, Islamic studies, maths, science, English, social studies and moral education take most of them. One period is labelled Artificial Intelligence and Technology.
That single period is her class's entire formal exposure to the subject this week, in a country that made AI a compulsory school subject from kindergarten to Grade 12 and expects the technology to contribute AED 335 billion to its economy by 2031.
The curriculum is not the hard part. The United Arab Emirates built one, approved it, extended it to private schools on 2 September 2026, and published a framework with seven content areas. The hard part is that a school week holds a fixed number of hours, a teacher has a fixed number of working days, and both were fully allocated before AI arrived.
The problem is arithmetic, not pedagogy
Begin with the grid. Under the 2026-27 teaching plan, Cycle 1 pupils get 32 lessons a week and Cycles 2 and 3 get 36, per the ministry's standardised allocation for national-curriculum schools. Multiply 36 by a 40-week year and a secondary pupil sits through 1,440 periods in a single academic year.
Since the subject became standalone in September 2026, AI takes one of those periods per week in Cycle 3, which is 2.8% of contact time. In Cycles 1 and 2 the frequency set in 2025 was one lesson every two weeks, which works out at 1.4% of the 2026-27 timetable.
Those shares come from the phased rollout that began in the 2025-26 year, when AI sat inside the existing Computing, Creative Design and Innovation block rather than standing alone. From 2026-27 it stands alone under the name Artificial Intelligence and Technology, and the ministry has said it stays inside the approved timetable rather than becoming an extracurricular programme. That is the right call on principle. It also means the subject was inserted into a grid that had no empty cells.
Something had to give. The ministry's answer, announced with the 2026-27 framework, was to merge Design, Innovation and Computer Science into the new subject, so the hours came from an existing allocation rather than a new one. That solves the timetable. It creates a different problem, which is that the computing hours were already doing a job.
Then there is the assessment design. As of the 2026-27 year the AI subject carries no formal examination. Assessment is project-based, which puts the marking load outside the period allocation, in hours a teacher does not have. A subject that occupies 2.8% of the timetable and 100% of a project portfolio is not a 2.8% workload.
Table 1: The UAE school week before and after AI became a standalone subject
Item | Cycles 1 | Cycles 2 and 3 | What changed in 2026-27 |
Timetabled periods per week | 32 | 36 | Unchanged |
AI lessons per week | 0.5 | 1 | Moved out of the computing block |
Share of contact time | 1.4% | 2.8% | Now a named subject |
Formal examinations | None | None | Project assessment only |
Teachers delivering it, 2025-26 | 1,000 specialists | 1,000 specialists | 22,000 to be trained |
Source: Arabian Post on the September 2026 Cabinet decision 🔗, September 2026, and Edarabia on the 2025-26 rollout 🔗, 2025
Twenty-two thousand teachers, one bottleneck
The September 2026 Cabinet decision put a number on the training task: 22,000 teachers and education professionals to be equipped to use AI in teaching, assessment, curriculum analysis and lesson planning. The 2025-26 rollout had relied on about 1,000 specialist teachers.
Divide 22,000 by the 1,000 specialists hired for the 2025-26 year and the cascade ratio lands at 22 to one. Each specialist is, in effect, expected to bring 22 colleagues up to a working standard. If that happens in a single session per colleague, the model is a briefing rather than training. If it happens in a proper sequence, the specialists stop teaching.
The ministry has not left this entirely to the cascade. In May 2026 it launched a National AI Upskilling Programme for Teachers with Hamdan Bin Mohammed Smart University, aimed at practical and ethical use in daily school work. Dubai runs a parallel track: the DP World Foundation, the Knowledge and Human Development Authority and MIT RAISE launched a multi-year AI literacy programme in February 2026, targeting about 80,500 students and 3,600 teachers by February 2030, focused on Grades 6 to 8 and embedded inside maths, science, computing, art, English and Arabic rather than taught as a separate subject.
Three programmes, three delivery models, one calendar. That is the constraint nobody has priced yet.
What the February 2026 rule removes
One rule changes the shape of the problem, and it is easy to miss because a different part of the ministry issued it.
In February 2026 the Ministry of Education published a manual on the safe and responsible use of AI in classrooms. It prohibits generative AI tools outright for students under 13, or below Year 7. Above that age, use is permitted under supervision, with a ban on entering any personal data about students, teachers or parents into an AI system, and a requirement that both teachers and pupils verify anything generated before using it.
Line that up against the curriculum structure published for 2026-27. Cycle 1 covers ages four to ten and Cycle 2 covers eleven to fourteen. Both sit below or across the threshold. Together they hold most of the roughly one million pupils enrolled in public schools that year.
So the youngest two cycles study artificial intelligence, its ethics, its data and its algorithms, without touching a generative model. That is defensible as pedagogy. It is expensive as logistics, because the teaching has to run on offline materials, printed activities and teacher-designed simulations, and each of those has to be built for Arabic and English separately.
The ban also removes the cheapest training route available in 2026. The fastest way to make 22,000 teachers comfortable with a tool is to let them use it on their own lesson plans. For anyone teaching below Year 7, that route is now partly closed inside the classroom, which pushes the practice into personal time or into the subset of staff teaching older grades.
The schools that were added in September
The 2 September 2026 decision did more than extend a subject. It extended the obligation to private schools, which in Dubai alone educate the large majority of children in a population that is roughly 88% expatriate.
The detail matters for costing. Private schools following the ministry curriculum have been teaching the subject since 2025-26. Private schools on British, American, IB or Indian curricula have the direction but not the timetable: as of October 2026 the start date and delivery format for those schools had not been published. Until it is, each group is building AI into computing lessons on its own budget and its own schedule.
That produces a split procurement market. GEMS Education operates more than 45 schools across the UAE in four curricula. Each of those curricula has a different inspection framework, a different teacher licensing route and a different definition of what counts as a computing lesson. A ministry-issued textbook does not map onto all four.
There is a staffing constraint layered on top. Every teacher in a Dubai private school needs a KHDA licence, and the process runs four to eight weeks end to end once documents are attested. Hiring a teacher from abroad, including attestation, work permit and visa, takes eight to twelveweeks, according to regional recruitment firms in 2026. Any plan that assumes a fast injection of AI-capable teachers runs into that queue before it runs into a budget.
The assessment nobody counted
One cost sits outside every budget line in the three paths, and it scales with the curriculum rather than with the training.
The AI subject carries no examination in 2026-27. It is assessed by project. Assume two projects per pupil per year, which is conservative for a subject with seven content areas. Assume a class of 30. That is 60 project submissions for each of the 22,000 teachers, or 1.32 million assessments a year across the system.
Give each one eight minutes of teacher time for marking against a rubric. That is 176,000 teacher-hours, which at 1,440 annual contact hours per teacher works out at about 122 full-time teaching posts consumed by marking alone. Nobody has hired those 122 people in 2026.
The number is sensitive to the assumption, and that is the point of showing it. At four minutes per submission the load is 61 posts. At twelve minutes it is 183. Any of those numbers dwarfs the AED 5.5 million licence fee in path two and rivals the AED 42 million release cost in path one.
This is the real reason the assessment design matters more than the platform choice. Project-based assessment is the correct pedagogy for a subject without a right answer. It is also the most expensive form of assessment to run at scale, and in 2026-27 it is being layered onto a timetable with no slack and a marking load with no allocation.
Path one: buy the hours back
The first route is the one the ministry is already on, executed properly. Take teachers out of class, pay for cover, and run a structured sequence rather than a briefing.
The arithmetic is not complicated. Give each of the 22,000 teachers three days of release across the 2026-27 year. That is 66,000 teacher-days, or about 367 teacher-years of classroom time at 180 school days a year. Cover cost at AED 600 per day, built from a public-school daily rate of AED 364 to AED 545 plus overhead, comes to about AED 39.6 million for the year.
Add delivery. A trained facilitator running a cohort of 25 teachers for three days needs 880 facilitator-days across 22,000 teachers. At AED 2,500 per facilitator-day in 2026 prices, including materials and travel, that is another AED 2.2 million.
Total: roughly AED 42 million, one-off, with the caveat that 367 teacher-years of lost contact time is a real cost booked against pupils rather than against a budget line.
The advantage is that it works. Release time is the only training model with a reliable record of changing classroom practice. The disadvantage is political. In a system that publishes school inspection results, 66,000 teacher-days is a visible number, and every day a teacher spends in a training room is a day a class goes untaught.
Path two: put the training inside the product
The second route stops buying training as a separate event and buys a platform whose onboarding is the training. Adaptive learning products, lesson-planning assistants and assessment tools ship with guided setup, sample lessons and in-product coaching, and the teacher learns the tool while using it with a class.
The unit economics are far kinder. Commercial licensing for K-12 platforms in the Gulf ran at roughly AED 250 to AED 400 per teacher per year at this volume in 2026. Multiply by 22,000 and the annual bill is AED 5.5 million to AED 8.8 million, or about AED 28 million to AED 44 million across a five-year licence.
That looks like a bargain against path one until the hidden lines appear. Devices and connectivity are not in the licence. Arabic-language quality varies enormously between vendors, and a product that performs well in English often degrades into translation in Arabic. The February 2026 rule requires age gating, so the same platform has to be configured differently for Cycle 1, Cycle 2 and Cycle 3, tripling the integration work.
The deeper cost is lock-in. A five-year licence at AED 44 million buys a workflow teachers will not want to abandon, which is exactly the position a procurement officer wants to be in at renewal.
Path three: hire your way out
The third route stops training generalists and hires specialists. The country already employed 1,000 of them in the 2025-26 year. Add 1,000 more and the cascade ratio halves from 22 to one down to 11 to one.
The cost per hire is knowable. Public school teachers in the UAE earned AED 8,000 to AED 12,000 a month in 2026, tax free, with housing and an annual return flight in the package. Call the fully loaded cost AED 136,000 to AED 184,000 a year including housing, flights, insurance and end-of-service gratuity. Two thousand specialists cost AED 272 million to AED 368 million a year, recurring, forever.
That is six to nine times the one-off cost of path one, every year, and it does not solve the problem it appears to solve. A specialist teaching one period a week to 30 classes in 2026-27 is still one person. The bottleneck is not the number of people who know AI. It is the number of hours in a week.
Table 2: Three routes to the same 22,000 teachers, costed at 2026 prices
Route | Unit cost | Volume | Total | Recurring? |
Release time plus facilitators | AED 600 per cover day | 66,000 teacher-days | AED 42 million | No |
Platform with embedded training | AED 250 to 400 per teacher per year | 22,000 licences | AED 5.5 to 8.8 million a year | Yes |
Specialist hiring, 1,000 more | AED 136,000 to 184,000 per post | 1,000 posts | AED 136 to 184 million a year | Yes, permanently |
Source: UAE teacher salary benchmarks 2026 🔗, and Ministry of Education curriculum announcements 🔗, September 2026. Cover and licence rates are working estimates, not published tariffs
Three approaches, three bills
Set the three side by side and the trade is clear. Path one buys capability and spends contact time. Path two buys contact time and spends capability. Path three buys neither and spends money.
Path one is the only route that changes what a teacher can do in front of a class, because it is the only one that forces practice with feedback. It is also the only one that removes 367 teacher-years from the system in a country where inspection ratings are public.
Path two is the only route that scales without a calendar fight, because the training happens inside the working day. It is also the route most likely to produce teachers who can operate one product and explain no principle.
Path three is the only route that works immediately, because a specialist arrives trained. It is also the most expensive per unit of actual change, because the constraint was never headcount.
The workable answer is a blend, and the emirates have assembled one by accident. The ministry runs path one at national scale. Dubai runs a version of path two through the MIT RAISE partnership, embedded in existing subjects. The 1,000 specialists represent path three at pilot scale. None of the three has been costed against the other two in public, which is why each looks affordable inside its own budget line.
The fix that survives a timetable
If the goal is a teacher who can use AI without being handed extra hours, the only intervention that fits the grid changes what happens inside the 36 periods rather than adding to them.
Three moves, in order of cost.
· Move the assessment, not the teaching. Project-based marking is where the hours actually go in 2026-27. A shared rubric and a moderation pool across schools cut marking time per teacher without touching the timetable.
· Put the practice inside a subject that already owns a period. Arabic, science and social studies all have weekly hours. An AI task inside a science period trains the teacher in the tool and covers the science content at the same time.
· Train the 1,000 specialists to coach rather than to teach. A specialist who spends three periods a week coaching colleagues in their own classrooms reaches more teachers than one who spends 20 periods a week in front of pupils.
None of those requires a new budget line, which is both their strength and their weakness. Nothing that requires no budget also requires an owner, and programmes without an owner do not survive a change of minister.
The number worth putting on a wall is not AED 335 billion. It is 367 teacher-years, the classroom time that path one consumes in 2026-27. Divide the projected AI contribution by the teaching hours it costs to produce a workforce that can use the technology, and the ratio either justifies the programme or it does not. That is a decision a ministry can make. A headline cannot make it for them.
Calculations
Data as of October 2026. Period allocations come from the Ministry of Education's standardised teaching plan for national-curriculum schools in the 2026-27 year: 32 weekly lessons for Cycle 1, 36 for Cycles 2 and 3. The 40-week school year is the assumption used here for annualising. The actual instructional calendar in 2026-27 is shorter once exam and holiday weeks are removed, which makes the contact-time shares in Table 1 slightly generous to the programme.
Release-day costing uses AED 600 per cover day, built from the 2026 public-school salary band of AED 8,000 to AED 12,000 a month divided by 22 working days, plus roughly 20% for overhead. Facilitator cost uses AED 2,500 per day, an estimate based on regional training day rates rather than a published tariff.
Platform pricing at AED 250 to AED 400 per teacher per year is a working estimate for volume licensing in the Gulf K-12 market in 2026. It excludes devices, connectivity and integration. The specialist cost of AED 136,000 to AED 184,000 per post adds housing, flights, medical insurance and gratuity to the tax-free salary band.
Three caveats. The 22,000 figure covers teachers and education professionals rather than classroom teachers alone, so the training population is smaller than the headline implies but the release-day cost per classroom hour is higher. Some of the 1,000 specialist teachers may already be counted inside the 22,000, which would double-count part of the denominator. And the AED 335 billion figure is a national economic projection to 2031, not a return on the education budget, so dividing it by any education input produces a ratio with no meaning.
Sources
· Arabian Post, UAE extends AI curriculum to every school 🔗, Cabinet decision and 22,000 teacher target, September 2026
· Edarabia, over 1,000 teachers to teach AI curriculum 🔗, 2025-26 rollout details, 2025
· Analytics Insight, how the UAE is preparing students for an AI future 🔗, Dubai and MBZUAI programmes, 12 August 2026
· UAE Signal, teacher salary benchmarks 2026 🔗, salary bands and package structure, 2026
· ReapHR, hiring teachers in the UAE 🔗, licensing and visa timelines, 2026
This article was prepared by the expert team at Global Infinity Labs, drawing on insights and data provided through its MCP data services.
About Global Infinity Labs
Global Infinity Labs is an AI innovation company specializing in MCP data services.
Email: support@globalinfinitylabs.com
Copyright 2026 ACN Newswire. All rights reserved. www.acnnewswire.com
© 2026 ACN Newswire. All rights reserved. A division of Asia Corporate News Network.
