Who’s doing the thinking? What the IB’s AI design principles mean for schools

Listen on Spotify | Listen on Apple Podcasts | Listen on Podbean | Listen on Youtube

About this episode

The International Baccalaureate (IB) has published its first set of AI design principles to guide how artificial intelligence is used across its ecosystem. In this episode of IB Unplugged, Head of AI at the IB and returning guest Clara Hawking takes listeners inside how the principles were created and what they ask of schools.

Episode chapters

4:20 How the AI conversation has matured since the “tools moment”
8:27 Anchoring the principles to the IB learner profile
22:27 AI for human flourishing
28:39 Deepening learning: who’s doing the thinking
36:35 Educator agency and the human in the loop
41:04 Red lines: the decisions that must stay human
45:13 AI policy versus AI governance culture
52:03 A practical first step for the new year

From a “tools moment” to a question of governance

The conversation about AI in education has grown up. Clara recalls how the early phase, through 2023 and 2024, was mostly about tools: learning to prompt, building AI literacy, experimenting. Schools have since moved from experimentation toward implementation, and with it a harder set of questions.

Those questions are increasingly about governance. Leaders are asking how AI should be run inside a school, parents are entering the conversation, regulators are arriving in earnest, and concerns about assessment have sharpened as it becomes clear that cognitive offloading is real. The shift Clara sees is from “can we use AI” to “how should we use it.”

Anchored to the learner profile, not the technology

What makes the IB’s approach distinctive, Clara explains, is where it starts. The working group deliberately refused to treat this as an IT project. Instead of asking how the IB should fit into the technology, it asked how technology could serve the IB learner profile.

Strong external frameworks from the OECD, UNESCO, the Brookings Institution, and the European Commission were drawn on, but as inputs that serve the learner profile rather than replace it. Clara credits that shared starting point for making it possible for the group to reach a consensus: once everyone agreed on the anchor, most of the disagreement fell away.

The five principles, and what they ask of schools

The principles were shared with the IB community, tested through surveys and feedback at global conferences, and refined accordingly.

AI for human flourishing

The first principle puts human development first. Clara points out that AI has too often been treated as the goal in itself, when the goal is the human being. Emotional, social, cognitive, and ethical development stay central, and technology has to fit around that purpose.

“Human flourishing has to remain the objective and technology has to fit into that objective, not the other way around.”

Deepening learning

The second principle asks whether a tool deepens learning through inquiry, critical thinking, and creativity, or simply does the thinking for the student. Clara’s test for this is simple: is the learner thinking more, or less, than before they used the tool? She reaffirms that learning sometimes requires struggle, and that resilience and curiosity are built when students wrestle with difficult ideas rather than offload them.

“It’s really a question about who’s doing the thinking.”

Educator agency

The third principle, framed as guided by capability and grounded in responsibility, drew the most feedback from the community. At its core is the stipulation that the teacher remains the decision-maker, with a human always between the student and any AI output. Clara is emphatic that this includes the freedom to stay analog: a teacher may consciously choose not to use AI in a subject because the skills being developed are better built another way. Supporting professional judgment, rather than overriding it, is where AI earns its place.

Safety and transparency

The fourth principle addresses boundaries. Some decisions, Clara argues, must stay human: high-stakes grading and progression should never be fully automated, child safety cannot be compromised, and systems that undermine learner agency have no place. She also points to an emerging red line, which she expects to become more defined over time, regarding opacity – if systems cannot be understood meaningfully by the users, it is impossible for decisions to stay genuinely human.

Continuous adaptation

The fifth principle is characteristically IB: adapt in an evidence-led way as the technology changes. Consent given to one model is not consent to its successor, so review has to be ongoing rather than a one-time sign-off.

Policy is a document; governance is a practice

Clara draws a clear line between having an AI policy and having an AI governance culture. A policy on its own, she suggests, can become a dead document, produced once and filed away, like a dress code no one revisits until something goes wrong. Governance is different. Drawing a parallel with wellbeing, she advocates accountability, oversight, and ongoing review being woven into the institution’s daily life, so that AI is thought about in every policy, from hiring to field trips, rather than parked in a single document.

A practical first step for the new year

For schools wondering where to begin, Clara has concrete suggestions. Convene a working group, but a broad one: not only IT and the early adopters, but someone from wellbeing and safeguarding, students who are not necessarily tech-savvy or high-performing, and a parent representative where possible. Use it first to map out the existing AI use already happening across the school, much of it unseen, through an anonymous survey. Then choose one high-risk area and work on that, before reaching for new tools. In her framing, you cannot govern what you have not yet mapped.

About Clara Hawking

Clara Hawking is Head of AI at the International Baccalaureate, and a globally recognized leader in AI governance and education. She was one of a small group of advisors who shaped the IB’s first AI Design Principles. She is also Co-Founder and Executive Director of Kompass Education, was previously Head of AI for a network of more than 70 schools across four continents, and won the Tech Women 2024 Global Achievement Award for her advocacy for women and girls in technology. She is a sought-after speaker and a trilingual parent of two.

You can hear the full conversation with Clara Hawking in this episode of IB Unplugged. Listen and subscribe wherever you get your podcasts.

Podcast