15/09/2026  •  7 min read

academic integrity | AI in education | critical thinking | curiosity | IB DP | metacognition

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

About this episode

Tim Lovatt is Head of Digital Learning at UWC South East Asia and a physics teacher who leads whole-school work on AI and academic integrity. In this episode of IB Unplugged he makes a straightforward case: AI is neither good nor bad for curiosity on its own, and what decides the outcome is how intentionally students and teachers use it, held together by a clear school culture.

Episode chapters

• 7:12 The same tool can squash curiosity or amplify it
• 11:17 Making AI increase the friction of learning
• 12:16 Overwhelm, deadline stacking and the real roots of cheating
• 14:21 Teaching metacognition from primary through the DP
• 15:01 Redesigning assessment around the process of understanding
• 23:41 Anchoring AI decisions to a clear school mission
• 34:06 The line of best fit and staying in charge of the machine
• 41:39 Winning teachers over by giving them room to explore

From a “tools moment” to a question of governance

Walk into almost any IB DP classroom and you will find students who can get a fluent answer to almost any question in seconds. That access is not going away, and it changes what teachers are working with. The real question about using AI in the classroom is no longer whether students will reach for it. It is what that reach does to their thinking.

Tim Lovatt puts it plainly: the relationship between AI and curiosity is tense, and it can go either way. Ask the same tool for a quick, mostly-correct answer and a student’s instinct to wonder gets switched off without them noticing. Ask it to challenge a preconception or surface perspectives they would never have reached alone, and that instinct grows. The difference is the interaction, and the self-awareness behind it.

Make AI ask the questions

For Tim, the most useful shift a student can make is to stop treating AI as an answer machine. If you are only asking it questions, he tells his students, you are using it wrong. What should happen is closer to the reverse: the tool asks the questions, and the student does the thinking. Flip the narrative that way and the instinct to wonder becomes reflective, turning back on what a student thinks they already know.

“If you want to learn something, AI should be increasing the friction for learning, not decreasing it.”

Most consumer AI is built to do the opposite, to speed things up and smooth things over. That is fine for organizing a calendar or handling a task a student has already mastered. For learning something new, the friction is the point. Curiosity is harder to hold onto when answers come this easily, which makes the instinct to keep wondering matter even more.

Cheating is usually a symptom of overwhelm

It is tempting to read AI shortcuts as laziness. Tim pushes back on that.

“Teenagers don’t cheat because they’re lazy. Teenagers cheat because they have 12 different deadlines.”

In a competitive international-school context, a student might move from class through tutoring, music, and sport before starting homework late at night, and when the goal is immediate academic survival the shortcut is right there. The point for teachers is that integrity is downstream of design. If the goal on offer is a grade by tomorrow, AI will help students reach it; if it is genuine understanding, the calculation changes. Framing the work around learning rather than compliance does more for academic integrity than any detection tool.

Teach the thinking, then redesign the assessment

Helping students use AI well starts earlier than the DP. Tim argues for teaching metacognition explicitly, from primary school upward, so that by the time students reach the Diploma Programme they understand how their own learning works and can make deliberate choices about it. Thinking routines and reflective habits give them a way to interrogate what they already believe and to notice what they still want to know.

Assessment has to move with this. For too long, he suggests, assessment has asked how much a student can recall from a two-year course in a single exam. The IB is conceptual and dynamic, yet it still largely measures the understanding a student ends with rather than the process they went through to build it. Many teachers at his school have started redesigning tasks around that process, through oral defenses, reflective examination of a student’s own writing, and exhibition-style arguments. Such tasks are hard to outsource: the evidence of learning is the student’s own reasoning, shown live.

Anchor it to the school’s mission

Underneath all of this sits a question about purpose. Tim’s clearest advice is that a school needs organization-wide clarity on why it does what it does, because that sense of purpose is what students draw on when they decide how to use a tool. At UWC South East Asia the mission is explicit and lived:

“Our role is to make education a force to unite people, nations, and cultures for peace and a sustainable future.”

It runs through strategy documents, new initiatives, and the questions teachers put to students. Conversations about AI at his school are values-based rather than rule-based: instead of a list of permitted and forbidden uses, teachers are transparent about what a task is for and which skill it is meant to build. If AI would do the very thing the student is supposed to be learning, that is the line. Students who understand the reason behind a task make better calls than students following a rule they cannot see the point of.

Using AI in the classroom to deepen thinking

Tim’s favorite example is deliberately low-tech. Teaching graphing in physics, he had students plot a line of best fit by hand. When they asked why the expensive machine in front of them could not just do it, he had an answer ready: when you draw that line, you decide where it goes, and that judgment is the thing being measured. “You are in charge. You are the scientist.” Make those decisions a few times by hand, and later, when software does the plotting, you have the critical lens to notice when it has got it wrong.

Used this way, AI becomes an amplifier for thinking rather than a substitute for it. Tim describes design students stress-testing a prototype against AI-generated personas before meeting a real person with lived experience, so the real conversation lands with far more depth. The common thread is low-stakes practice: a safe space to try, fail, and try again, which is where a lot of real learning happens. He makes the same case for teachers: rather than mandating which tasks AI may touch, he gives staff unstructured time to solve their own workflow problems and build their own confidence.

About Tim Lovatt

Tim Lovatt is Head of Digital Learning at UWC South East Asia, where he leads whole-school work on AI integration, edtech strategy, and academic integrity across K–12. A physics teacher by background, he focuses on metacognition, assessment design, and building school cultures where students learn to use technology intentionally. His work spans classroom practice and the wider question of how schools keep curiosity alive.

You can hear the full conversation with Tim Lovatt, including the two parent meetings that captured the whole screens debate in half an hour and why he thinks productive, low-stakes failure belongs in every classroom, in this episode of IB Unplugged. Listen and subscribe wherever you get your podcasts.

Podcast