What are AI Sprints?
Live, interactive conversations with trusted voices about AI in education. Each session focuses on practical, ethical use in classrooms and leadership, with the audience shaping the discussion.
Season 2 starts on 24 September. Save your place here.
If you missed Season 1: Catch up on the highlights.
Episode 1 — Darren Coxon and Priya Lakhani
Theme: Intentionality.
How leaders talk about AI matters. If adults are fearful, pupils will mirror it. Priya argued the “AI train” has left the station, so schools should lean into deep thinking and deep learning, not resist. She warned that rushing to over-regulate can stifle innovation and push start-ups away. The challenge for education is clear: when harm happens, how do we preserve trust while continuing to innovate?
Episode 2 — Daisy Christodoulou and Lord Jim Knight
Theme: Assessment and the “Black Box”.
Daisy asked whether we should put AI back in the Black Box to acknowledge limitations in how systems make decisions. She highlighted risks like students gaming models that reward length over quality. Jim pressed for transparency so teachers and pupils can trust outputs and respect copyright.
Evidence: Daisy’s team at No More Marking studied 5,251 Year 7 pupils from 44 schools and compared teacher and AI judgments. Agreement was 81 percent, with most of the 19 percent disagreement linked to human bias. Type of error matters: position bias can flip AI judgments 10 to 25 percent of the time.
Looking ahead: Jim suggested exams may shift toward tools like voice-to-text but warned about the digital divide. He echoed Rose Luckin’s advice to think now and act slowly. Daisy reminded us that knowledge is a skill, not a bolt-on.
Episode 3 — Laura Knight and Jon Chippendall
Theme: Thinking with care.
Laura contrasted “move fast and break things” with a purposeful approach. Both guests argued for a “magic middle” that balances opportunity and risk.
Classroom language:
-
Augmented ideation uses AI as a critical friend to spark ideas.
-
AI slop refers to shiny but shallow outputs.
-
Glazing is the flattering layer that can make AI feel better than it is.
The warning is not to offload thinking to AI. Keep the human in control of the loop. Jon noted AI can be “scarily good and confidently wrong”. Both highlighted strong potential for SEND support when used well.
Episode 4 — Bukky Yusuf and Al Kingsley MBE
Theme: Equity and missing voices.
Who is AI designed to care about? Technology itself does not care, so teachers must. Al described a third digital divide based on skills and understanding, not just access. Both pointed to gains for neurodiverse learners and pupils with EAL when tools are accessible and safe.
Critical literacy: pupils use AI in school and out of school. We need to teach what to be critical of. Live demos showed how bias, hallucinations and anthropomorphism creep in. Try asking any LLM for the “top 10 scientists” and notice how narrow the list can be.
Episode 5 — Jon Dolman and David Curran
Theme: Practical classroom use.
Jon integrated AI into the feedback loop for essay writing. He modelled a first paragraph, pupils wrote under timed conditions, then used AI to generate and critique alternatives. This built metacognition as pupils acted as coaches for each other and the tool.
David used MindJoy chatbots to personalise learning, especially for pupils with dyslexia. Domain-specific bots supported subjects and safeguarding. With high moderation settings, the system flagged “sad” language in a creative task, prompting timely teacher support.
What Season 1 taught us
-
Be intentional. The tone we set about AI shapes pupil attitudes and behaviours.
-
Use evidence and build trust. AI can align closely with teacher judgement, but biases and error types still matter.
-
Keep the human in control. Avoid intellectual offloading. Teach pupils to evaluate outputs.
-
Prioritise equity. Address access, skills and understanding so AI reduces rather than widens gaps.
-
Focus on practice. Strong use cases combine pedagogy, safety and real classroom value, especially for SEND.
Thank you and what is next
Season 1 showed AI’s promise and its pitfalls. It proved that careful thinking, practical experimentation and inclusive design can move us forward.
Thanks to our sponsors: Brainfreeze, TeachMate AI and Quizizz
Season 2 starts on 24 September. Save your place here: https://forms.office.com/e/LSgh5MBUH3