Of the sessions I attended, the one I found most useful was on Cogniti, a platform that allows teachers to develop AI agents tailored to student learning. This challenged me to design an AI “coach” to guide students through their IGCSE Global Perspectives Team Research Project. The idea is not that AI gives students the answers, but that it poses questions, prompts reflection, and scaffolds their thinking—from developing a research question to writing their reflective paper.

My first trial, where I posed as a student, revealed an important lesson: the AI’s feedback was too much, too fast. For this to be useful, it needs to “chunk” feedback into smaller, more manageable steps that students can process and respond to in their own time. I have a little more work to do on it! This is where AI shows both its promise and its challenge: the technology is powerful, but how we shape its interaction with learners will determine whether it deepens or diminishes critical thinking.

Other sessions broadened the picture. We explored AI as a research tool, experimented with the craft of prompt writing, and even trialled AI’s capacity to mark assessments. In this last workshop, assessments already double-marked by the teaching team were run through Microsoft Copilot. The results were striking. Copilot’s marks closely aligned with the teachers’, and the feedback it produced was both insightful and constructive. While no one suggested that AI could or should replace the teacher’s judgment, nor that there is going to be a policy change regarding who does marking in the future, the experiment highlighted its potential to complement marking by providing richer, faster feedback.

Perhaps the greatest value of the conference, however, came from the conversations that happened in between the formal sessions. Around coffee cups and corridor chats, colleagues wrestled with AI’s ethical and educational implications. One colleague expressed deep concern about whether AI tools draw on information that may be pirated or misused – a reminder that questions of intellectual property remain unresolved. Others reflected on AI’s potential to sharpen rather than blunt students’ critical thinking skills, provided we frame it carefully and encourage thoughtful use.

The biggest takeaway for me from the conference was not whether AI will shape education. It already is. Nor was it simply the reminder that educators must shape its use wisely; that much is clear. What struck me most was how often our conversations stay focused on immediate questions: Will AI change assessment? Will it speed up feedback? Will it help or harm? These are important, but they are too narrow. AI’s rapid growth is reshaping the world more profoundly—and more quickly—than many of us realise. The real takeaway, though, is that we need to look beyond these immediate concerns.

The comments of the Barker Journey students show the importance of this. Students are asking how to use AI responsibly and with integrity. Their questions differ from ours: while we often focus on tools and systems, they are already thinking about the world they want to live in. They see AI as part of that world, and they want the skills to shape its use for the common good. For educators, then, the task is clear. We are responsible for preparing the next generation to live and lead in a world where AI will be ever-present. Our role is not to prescribe the world they must build, but to help create the conditions for the world they wish to inherit.