The IB Is Changing What Counts As Learning

In Short:

• The IB’s new Systems Transformation Pathway replaces two Standard Level subjects with a 300-hour course inside the Diploma Programme.

• This matters because AI has made polished take-home work weaker as stand-alone evidence of student thinking.

• Many systems are responding with AI literacy, platforms, responsible-use rules, or assessment policies. IB is changing the shape of the qualification itself.

• The next question is whether IB can use AI to help train teachers, support schools, and keep the reform current.

A student submits a polished essay. The argument is clear, the grammar is smooth, and the references look plausible.

What can the teacher still know about how the student thought?

That question existed before ChatGPT, but AI has made it harder to avoid. When a fluent answer is cheap, the thinking behind it matters more.

The International Baccalaureate saw part of this early. Its position was not to pretend AI could be banned. The IB said it would not ban AI software, but students had to credit AI-generated material and could not submit AI-produced work as their own. Its broader AI guidance also keeps assessment human-led, with AI explored only as a possible quality-control tool.

That context makes the IB’s new Systems Transformation Pathway worth watching. It’s the result of a broader 16+ review of the Diploma Programme and Career-related Programme. But AI gives the reform sharper relevance: if technology changes what students can produce, schools need to rethink what students should prove.

The Systems Transformation Pathway can replace two Standard Level subjects in the Diploma Programme. IB says the pilot is intended to expand in 2028, with mainstream first teaching planned for 2030.

In plain English, this is not an after-school project. The Upper Canada College subject brief describes a 300-hour course taken alongside three Higher Level subjects, one remaining Standard Level subject, and the IB core. Students still work towards the full Diploma, but some of what counts for the Diploma shifts from separate subject coverage to systems thinking, design thinking, projects, portfolios, and real-world work.

For a qualification body, that is gutsy. Every change touches university admissions, parent expectations, school timetables, teacher workload, and student futures. IB is making room inside the credential for a different kind of learning.

The point is not that essays are dead. (They are not.)

The better point is that a final artefact no longer tells us enough on its own. A student may still need to write well, argue clearly, and research carefully. But we also need to know how they framed the problem, what help they used, how they revised their thinking, and whether they can explain their choices.

That is why detection is the wrong place to centre the conversation. Research by Weber-Wulff and colleagues found AI text detectors were not accurate or reliable enough for academic settings, and OpenAI withdrew its own classifier because of its low rate of accuracy. Academic integrity still matters, but schools should not build the future of assessment around catching students after the fact.

Higher education quality agencies and regulators are asking a better question: what evidence shows that learning actually happened? The UK’s Quality Assurance Agency has urged universities to rethink assessment for the ChatGPT era, including more authentic and oral assessment. Australia’s higher education regulator has similarly argued for assessment that captures authentic demonstrations of learning.

In that light, the IB pilot is even bolder and more interesting. Case studies, projects, portfolios, collaborative work, and oral defence are not AI-proof, but they ask students to show more of the path: how they worked with others, made tradeoffs, changed course, and defended their judgement.

How have other systems been responding? UNESCO has published AI competency frameworks for students and teachers, giving schools a language for AI literacy, ethics, and human-centred use. Singapore’s EdTech Masterplan 2030 is more infrastructure-led, with AI tools built into its national Student Learning Space. Australia has a national framework for generative AI in schools, focused on teaching and learning, wellbeing, privacy, transparency, fairness, and accountability.

New Zealand’s guidance is especially clear on assessment authenticity: schools must include acceptable GenAI use in assessment policies, while GenAI is not permitted in NCEA external assessment. Wales offers another route, connecting AI to digital competence, safeguarding, professional learning, and a purpose-led curriculum.

These are all necessary responses. Schools need AI literacy. They need privacy rules. They need platform decisions. They need clear policies for assessment. Complementary to all of that, IB’s move is asking whether learning time inside a global qualification should be reorganised around complex systems, action, reflection, and visible evidence of process.

Curriculum reform often fails between the vision and the classroom. The OECD has written about this implementation gap, including the need for teacher capacity, stakeholder buy-in, and protection against reform fatigue. IB knows this too. In a TES interview, Director General Olli-Pekka Heinonen named teacher preparation and university understanding as major challenges before 2030.

I, for one, am curious to see how the IBO proposes to use AI in teacher training. IB could use AI to help teachers compare sample portfolios, test project briefs before they reach students, translate early pilot lessons into usable guidance, and explain the pathway to parents and universities. Human judgement would still need to lead, especially in assessment. But AI could make the support system around the reform stronger.

That is why this pilot is worth watching. IB has enough trust from schools and universities to attempt something many individual schools could not do alone. If it treats Systems Transformation as a living reform, not a finished 2030 product, it could become one of the more serious models for education after ChatGPT.

For teachers and school leaders, the lesson is not “copy the IB”. It is simpler than that.

This term, choose one assignment where the final product carries too much weight. Add one visible-thinking checkpoint: a process log, oral defence, conference, AI-use reflection, or portfolio extract. You do not need to redesign the Diploma Programme to start moving assessment back towards thinking you can actually see.