IB and IGCSE tutors for Green Community families

Almost nobody sets out to cheat. The breaches that get students into trouble in Dubai schools are mostly produced by pressure and poor technique rather than by intent. A student with four unfinished components and three weeks left makes a decision at midnight that they would not make in September. A student taking notes without separating quotation from paraphrase ends up with sentences in a draft that are not theirs and cannot say where they came from. A student whose Arabic is weak runs a paragraph through a translation tool and submits it without thinking of it as dishonest.

The structural cause is the one described at the top of this page. The independent components are substantial, the supervision is deliberately rationed, and the deadlines cluster. A parent who understands that is in a position to intervene early, which is the only intervention that works. A parent who discovers it in the final fortnight is choosing between a poor submission and a dishonest one, and a market exists to make the second option easy.

The second cause is that nobody states the rules plainly to the student. Schools publish academic-honesty policies, and most students have signed one without reading it. They know plagiarism is wrong. They are far less clear about collusion, about submitting the same work twice, about whether a parent may proofread, about whether a tutor may suggest a structure, and about where a language tool sits. Those are the grey areas where real cases arise.

This page sets the boundary out explicitly because that is more useful than a reassurance. A student who knows exactly what is permitted works more confidently, asks better questions of their supervisor, and does not reach the midnight decision in the first place. It also makes a tutor's role clearer: the help that is legitimate is substantial, and knowing its shape is what stops a family buying the wrong thing in a panic.

School and board policies tend to describe four or five categories rather than one. Plagiarism is representing someone else's words, ideas or data as your own, including from a website, a textbook, a sibling's old work or a generative tool, and including when it is accidental. Collusion is supporting another candidate in presenting work that is not their own, which catches the student who shares a completed answer as well as the one who copies it.

Duplication of work means submitting the same piece for two different assessments, which students do innocently when a topic suits both an essay and an investigation. Fabrication of data covers inventing results or adjusting them to fit a hypothesis, which happens quietly in science investigations where an experiment went wrong and the deadline did not move. Misconduct during an examination covers unauthorised material and disruption, and is a separate matter handled by invigilation rules.

The consequences are set by the board and the school rather than by any tutor, and they are proportionate to the component. A finding against a Diploma component can mean no grade awarded in that subject, which can mean no diploma. The point worth making to a sixteen-year-old is not the severity but the asymmetry: the operator who sold the work faces nothing, and the candidate faces the outcome.

What protects a student is evidence of process. A dated process journal, draft versions saved as they were written, notes that record where each source came from, and a supervisor who has seen the work develop. Students who keep those things are rarely in difficulty even when a similarity report flags something, because they can show how the work came to exist. Students who produce a finished document from nowhere have no defence available to them.

When a component is submitted, the supervisor confirms it is the candidate's own work. That confirmation is not a formality and it is not based on trust alone. Supervisors are expected to have seen the work develop, which is why schools require meetings at specified points, why a Personal Project is assessed partly on a journal kept during the project, and why a supervisor who has never seen a draft is in a difficult position when something looks wrong.

The checks available are more varied than students assume. Similarity-detection software compares a submission against published sources and against other submissions. Supervisors compare a written component against the student's classwork, and a sudden change in vocabulary, syntax or confidence is visible to anyone who has marked that student for a year. Where doubt remains, a supervisor can question a student about their own work, and a candidate who cannot explain their own method or defend their own argument has answered the question.

That last mechanism is worth describing to a student explicitly, because it reframes the whole issue. The risk is not being caught by software. It is being unable to discuss your own work with the person who knows your writing best. A student who has done the work can talk about it for an hour; a student who has not cannot manage five minutes, and nobody needs a tool to tell the difference.

For a family, the practical implication is about record-keeping rather than suspicion. Keep the drafts. Keep the dated journal. Keep notes with sources attached. Our supervision sessions produce a written record of what was discussed and what the student agreed to do next, and we will release it to the school on request. That record exists to protect the student, and it is one of the reasons online supervision has an advantage over a conversation at a kitchen table.

This is the live question in every Dubai staff room and the one competitor pages avoid entirely. The honest position is that policies vary by school and are still moving, so the first and most important step is to read your own school's written policy and ask the department what it expects. What follows is a general framing, not a substitute for that policy, and where the two differ the school's rules govern.

The useful distinction is between a tool that helps a student think and a tool that produces the output being assessed. Using a search tool to find a paper, a calculator or spreadsheet to process data a student collected, a spellchecker, or a grammar prompt that flags an error for the student to fix, sits comfortably inside normal practice. Asking a model to draft a paragraph, to generate an argument, to summarise a source the student has not read, or to translate their own sentences into fluent English or Arabic, does not: the assessed output stops being theirs.

In between sits a genuine grey zone that needs declaration rather than avoidance. Using a model to explain a concept the student then writes about themselves, to suggest search terms, or to generate practice questions is usually acceptable and worth mentioning to a supervisor. Using one to check a reference format is low risk. The rule that holds across all of it: if a student could not reproduce the thinking without the tool, the tool produced the work.

Language subjects are where this bites hardest, and Arabic as an additional language is the specific case in this community. A student submitting translated text is submitting someone else's language, which is precisely what is being assessed. The legitimate route is slower and actually works: build vocabulary, practise the script, write badly in Arabic and improve it with a teacher. We will not supply translated text, and we will tell a student plainly that using a tool for it defeats the only purpose the subject has.

The legitimate scope of tutoring on an independently produced component is wide, and families who do not know that often either buy nothing or buy something they should not. A tutor may teach the underlying content, which is the bulk of the work. A tutor may read out the assessment criteria and explain what each strand actually rewards, which most students have never had done properly. A tutor may interrogate a research question until it is narrow enough to answer, which is the single highest-value hour available.

A tutor may show how a comparable published investigation was structured, as an example of form rather than content. A tutor may read one draft the student has written and respond with questions: what is your claim here, where is the evidence, why this method rather than that one, what would change your conclusion. A tutor may check that length, formatting and referencing meet the requirements in the current subject guide. All of that improves work without any of the tutor's words entering it.

What a tutor may not do is produce the output. Not write, not rewrite, not edit a draft into a finished state, not supply text, data, structure to be copied, or a model answer that could be submitted. Not translate. Not run a model and hand over the result. Not complete a section the student has left blank because they ran out of time. These are refusals rather than preferences, and we make them clear to the student in the first session so that nobody is negotiating them at midnight.

There is one further thing a good tutor does that is rarely mentioned: say no out loud, with a reason, in front of the student. A teenager who hears an adult decline to cross a line, and explain why in terms of the risk to them rather than to us, learns something more useful than a policy document. It also means that when the request comes from a stressed parent rather than from the student, the answer has already been given.

This market is visible to any parent searching in Dubai, and it presents itself as tutoring. The research behind these pages found a site advertising finished internal assessments at AED 3,750. Others are less explicit and no different in substance: they promise a completed component by a stated deadline, quote a fixed price per piece of work, and describe the service in language borrowed from editing and proofreading.

There are reliable signals. A fixed price per component rather than an hourly rate for teaching is the clearest, because teaching time cannot be priced per document. A promised turnaround measured in days. A grade attached to the offer. No named tutor, or a named tutor who never speaks to the student. An emphasis on confidentiality rather than on learning. And a refusal to put the scope of work in writing, which is the question that ends most of these conversations.

The second thing to look at is what happens to the student. An honest arrangement produces a student who can discuss their own work, because they did it. A purchased component produces a student who will be asked about their own method by a supervisor who has marked their writing for a year, and who cannot answer. The risk is not primarily detection software; it is the five minutes of conversation that any experienced teacher can conduct.

Parents occasionally ask whether a heavy proofread is a safe middle route. It is not, and it is the most common way an honest family ends up on the wrong side of a line. Correcting spelling is fine. Rewriting sentences for fluency changes the assessed output, especially in language subjects where expression is the thing being marked. If an adult's rewriting materially improved the text, the text is no longer the student's, whatever the intention was.

Supervision is priced differently from weekly teaching, and the difference is worth understanding before you compare quotes. Weekly teaching is an ongoing hourly arrangement. Supervision is a small number of sessions placed at specific points in a component's life: question design, method or structure review, a response to a draft the student has written, and a final check against criteria and formatting. Four hours placed well beat twelve bought in the final fortnight.

The figures below are Dubai market rates collected in October 2026, each attributed to whoever published it. They are market rates rather than an IB Gram price list, and they exist on this page because three of the six agency pages ranking for these searches publish no price at all. Use them to judge whether a quote you have received is normal, and to notice when a price is attached to a document rather than to time.

Note what the last row in the table is doing there. A price per finished component is not a tutoring rate at all; it is the price of a risk transferred to your child. It appears in this market at a figure comparable to several months of honest weekly teaching, which is the comparison worth making. The same money buys a term of work that improves a student and leaves no exposure.

One more number for context. KHDA confirmed in May 2026 that private school fees in Dubai would not rise for the 2026-27 academic year, under an economic incentives package worth AED 1.5 billion. Tuition is the movable part of the education budget this year, and for a Diploma household the highest-return use of it is a handful of supervision hours placed early in the first Diploma year rather than a block of hours bought under deadline pressure in the second.

A large share of honesty cases come from technique rather than from decisions. The most common mechanism is note-taking that does not distinguish between what was quoted, what was paraphrased and what the student thought themselves. Three weeks later the notes are the only record, the distinctions have gone, and sentences that came from a source arrive in a draft looking like the student's own. The student is not lying; they genuinely cannot tell any more.

The fix is a habit rather than a tool, and it takes one session to install. Quotations go in quotation marks with the source and page recorded at the moment they are written down. Paraphrase gets a marker and a source. The student's own thoughts get a different marker. Everything carries a reference from the first note onwards, not added at the end from memory. Students who work this way produce cleaner drafts and finish faster, because they are never hunting for where something came from.

Referencing style matters less than consistency. Schools and subjects differ, and the requirement in most cases is to use one recognised style properly throughout. The errors that cost marks are mechanical: a bibliography that does not match the in-text citations, sources listed that were never cited, a web page referenced with no access date, a data set used with no attribution at all. None of those is dishonest, and all of them are visible to a marker.

Similarity reports deserve a word, because students panic at them. A similarity percentage is not a plagiarism verdict; a well-referenced history investigation with substantial quotation can report high similarity and be entirely honest. The useful way to read a report is item by item, checking that each flagged passage is quoted and cited or genuinely the student's own phrasing. We teach students to do exactly that, and to use the report before submission as a check rather than after it as an accusation.

Send a component list rather than a request for a tutor: each subject and level, which independently produced components are outstanding, the school's own internal deadline for each, and the criteria document or task sheet the school issued. Add the most recent marked piece of work. A supervision plan is built backwards from those dates, and sending them turns the first hour into work rather than into an interview.

The first session establishes which of four situations you are in, because they need different responses: a question that is not yet answerable, data collected that will not support the required analysis, a draft that answers a different question from the one asked, or a component simply not started. The first three are fixable with a small number of hours. The fourth sometimes needs a conversation with the school about timing rather than a tutoring purchase, and we will say so.

The boundary conversation happens in that first session, with the student present. The tutor says what they will do, what they will not do, and why the line sits where it does in terms of the risk to the student. That removes the midnight negotiation entirely, which is the single most useful thing a provider can do for a Diploma household. We also keep a written record of each session and will release it to the school on request.

Three limits stated before any payment. We do not promise grades or points. We do not send tutors to homes in the UAE, as every lesson is one-to-one online. And we will not write, rewrite, translate, edit to completion or supply any part of an assessed component, however late the deadline is, however the request is phrased, and whoever is asking.