Where the IB stands on AI
The IB's public position has two parts, published three years apart. In 2023 it said it would not ban AI tools such as ChatGPT, likening them to spell checkers, translation software and calculators, and asked schools and students to use them ethically and effectively. The condition is transparency. Work produced in whole or in part by an AI tool is not counted as the student's own, so AI-generated text, images or graphs in a piece of assessed work have to be credited in the text and referenced. The academic integrity policy gained a guidance appendix on AI tools.
In March 2026 the IB published five draft design principles for schools adopting AI: caring and balanced, inquiry-driven, educator agency, safe and transparent, and continuously adapting. It also set a line, that high-stakes decisions about grading, progression and well-being must not be fully automated, and said it was piloting a chatbot on its Programme Resource Centre so that teachers can put questions to the curriculum library instead of searching it. These are drafts and pilots, so check ibo.org for the current wording before you build a policy on them.
Why inquiry makes AI both a risk and an opportunity
An IB education is built on students asking questions, finding and weighing sources, and reflecting on their own thinking. That is exactly the work a chatbot can appear to do for them. If a student hands the question to a tool and hands the answer to a teacher, the product may look fine and the learning has not happened.
The same features make AI useful when it is aimed at the process. A tool can play the sceptical reader who asks a student to defend a claim, offer a counter-example, or explain a concept a second way. The distinction the IB keeps returning to is between using a tool to think better and using it to avoid thinking, and a school's task is to make the first easy and the second visible.
What is likely to change in planning and teaching
Nobody can forecast this precisely, but the direction of travel in schools that are already using AI carefully is fairly clear.
- First drafts of unit ideas, questions and resources produced in minutes, then challenged by a planning team rather than copied
- Differentiation and language support prepared faster for multilingual classes, then checked by the teacher who knows the learners
- Feedback drafted from a teacher's notes, leaving more of a lesson for conversation
- A shift of effort from producing materials towards choosing, questioning and adapting them
- More attention on the quality of the questions students are asked, because answers are now cheap
What is likely to change in assessment
The IB's requirement to acknowledge AI use moves the weight from prohibition to evidence of process. Schools will lean harder on drafts, planning notes, in-class writing, oral explanation and reflection, because these show how a student arrived at work, and a teacher who knows a student's voice can tell a great deal from a short conversation.
AI detectors are not a substitute. Their output is a probability from a tool that errs in both directions, and a school that treats it as proof risks accusing a student who did nothing wrong. The stronger evidence is the trail of the work itself, which is why an IB academic integrity policy now has to describe process as well as rules.
What stays human
The line the IB has drawn is a good one for any school: consequential decisions about a child are made by people. Grading is a judgement against criteria, progression is a decision about a young person's path, and well-being is a matter for adults who know the child. A tool can assist with the preparation and can never own the decision.
The learner profile is the other anchor. Caring, reflective, principled and open-minded are not things a model has, and they are the reason a teacher's relationship with a class does not automate. The IB learner profile is a useful checklist for what a school is trying to protect.
What IB coordinators can do now
The most useful steps are unglamorous and can be finished within a term.
- Update the academic integrity policy so it describes acceptable AI use by task, and how students acknowledge it
- Agree subject-level expectations, because a language class and a mathematics class will not want the same rules
- Teach students how to cite and reference AI use, as an explicit skill rather than a warning
- Give teachers time to try the tools on real planning, and share what failed as well as what worked
- Check every tool against the school's privacy commitments before staff or students use it
- Tell families what is used, where and why, and keep a record of the decisions for the school's own review
How skodle helps
This section names our own product, so weigh it accordingly. Nothing here replaces the school's own policy and judgement.
skodle's AI sits inside the school platform. The curriculum copilot drafts unit ideas, central ideas and lines of inquiry from the school's own PYP and MYP units, and each draft stays pending until a person approves or rejects it, with higher-stakes topics needing a school leader; a second copilot turns the gap report into a prioritised list of fixes. Teachers can draft lesson plans, worksheets with answer keys, quizzes and feedback. For marking, the AI can draft a suggested score on short-answer, long-answer and rubric questions, but the suggestion becomes a mark only when a named teacher accepts it, and a blank or unreadable answer gets no score rather than a zero. Student tutors are created and switched on by staff, cannot be created by students, are instructed to teach rather than hand over finished work, are readable by staff, and pass a narrow set of distress phrases to a person. Every request is recorded, and staff can mark an answer helpful, unhelpful or wrong. You can test all of it on your own units in the free 30-day trial that every new school starts with: start the free trial. See skodle AI and the IB Engine.
The limits, plainly. The curriculum copilot reads only the twelve most recent PYP and MYP units and does not read curriculum maps, and approving a draft records a decision without updating the unit. A tutor's instruction to teach rather than answer is guidance to the model, not a guarantee, which is why staff can read the conversations. Requests, and any records a tool retrieves, are processed by an external model provider, and there is a daily usage limit that depends on the plan. skodle is not an AI detector and cannot tell you whether a student used AI, it does not predict grades, and it does not carry the official IB route for registration and coursework submission.
A realistic view of the next three years
Predictions in this area age quickly, so treat these as things to watch rather than things to plan around. School policy is likely to move further from prohibition towards disclosure and process evidence. Teachers' own use for drafting is likely to become ordinary, and the interesting question will be quality rather than whether it is allowed. Scrutiny of where student data goes is likely to increase, and the IB's own tools and guidance will keep changing.
The practical answer is a short review each term: what did staff use, what went wrong, what does the policy now need to say. A school that does that will not need to guess the future, because it will keep reading the present.
Questions to put to any AI tool used in an IB school
The IB's draft principles give a useful way to organise the questions.
- Caring and balanced: what does it do for a child who is struggling or distressed, and who sees it?
- Inquiry-driven: does it prompt thinking, or supply answers a student should have reached?
- Educator agency: who decides how it is used, and can a teacher switch it off?
- Safe and transparent: where does data go, and can we see what the tool did?
- Continuously adapting: how will we know whether it is helping, and what would make us stop?
Mistakes IB schools make with AI
The common ones are avoidable.
- Banning it outright, which pushes use out of sight and removes the chance to teach acknowledgement
- Relying on a detector as proof of misconduct
- Writing a policy once and never returning to it
- Letting each teacher improvise, so students meet five different rules in one day
- Adopting a tool because it is impressive in a demonstration and never testing it on the school's own work
