The questions worth asking: the environment and your data

Some of the most useful things we write come straight from the questions teachers and leaders ask us. When a school is weighing up whether to bring Starlight into its professional development, the people in the room are right to push. Hard questions are not an obstacle. They are the reason we keep sharpening how we explain what we do, and very often they are the reason a topic ends up here on the blog.

Two questions have come up repeatedly in recent weeks, and both deserve a clear answer. The first is about the environment. What is the real cost, in energy, water and land, of using AI in this way. The second is about data. Who can see it, where it lives, and what a line in our terms about possible future commercial use actually means.

We have written before about why trust is the real test for any AI coaching tool, and about the wider opportunities and risks of AI in schools. This post sits alongside those. It is written for the colleague who is not against Starlight in particular, but who is rightly cautious about AI in general.

The environmental question

The honest starting point is that AI is not free of environmental cost, and we would not pretend otherwise. What matters is the scale of that cost for this kind of tool, and the choices a provider makes around it.

The part of Starlight that runs continuously is the hosting: where lesson recordings, transcripts and reports are stored and processed. We host with Hetzner, in data centres in Germany and Finland. Those data centres run on 100% renewable electricity, hydropower in Germany since 2008 and wind power in Finland. They are cooled using outside air for the large majority of the year, which avoids the water-intensive evaporative cooling that many large data centres depend on. That speaks directly to the questions about both energy and water. You can read Hetzner's own account of how it powers and cools its data centres if you would like to verify any of this independently.

The AI step itself is lighter than people often assume. When a teacher uploads a lesson, Starlight turns the transcript into a written coaching report. That is a short, text-based task. It does not generate images or video, which are the genuinely energy-hungry uses of AI that sit behind most of the headlines. The footprint of producing a single report is modest, and smaller than the day-to-day footprint of many cloud tools a school already uses without a second thought. We have chosen green-powered, efficiently cooled infrastructure and a lightweight processing model on purpose, and we keep that under review as the technology and the evidence move on.

The data question

The second question usually starts with a line some readers have spotted in our terms, which mentions the possibility of using data for commercial purposes in future. It is a fair thing to query, so let us be precise about what it does and does not mean.

Starlight does not sell, and will never sell, teacher recordings, transcripts or coaching reports. Individual reports are private to the teacher who created them. School leaders only ever see anonymised, aggregated trends through the Constellation Dashboard. None of that is for sale.

The clause refers only to anonymised and aggregated data, stripped of anything that could identify a teacher, a pupil or a school. We use that for one core purpose, which is to improve the service for the schools that rely on it: understanding which coaching prompts are most useful, for example, or how the quality of feedback develops over time. It also allows us to contribute anonymised findings to educational research, which is a long-term commitment of ours. We are contractually bound never to attempt to re-identify anyone, and never to publish anything that identifies a school, teachers or students without written consent.

Location matters too. Our agreement prevents us from moving school data outside the UK or the European Economic Area unless equivalent legal safeguards are in place, so it could not simply be relocated to a country without those protections. Data stays encrypted in transit and at rest, and teachers can delete their recordings and reports at any time. For anyone who wants the full detail, our Data Protection Impact Assessment is available on request.

Keep asking

We would rather field these questions than avoid them. A teacher who interrogates where their data goes, or what the AI is really costing the planet, is exactly the kind of professional any good school is glad to employ. Those instincts make our product better, and they shape what we choose to write about here. So thank you to everyone who has raised them. Keep them coming.

If you would like to see how this works in practice, you can book a short demo.

Spark Insight with Starlight, and let the daylight in.

🎥 Subscribe to our channel here: https://www.youtube.com/@Star21-ai

🌐 Read more on our blog: www.coaching.software

💡 Explore the platform: www.starlightmentor.com

🐦 Follow us on X: @star21starlight

The Insight Engine is written by Adam Sturdee, co-founder of Starlight, the UK’s first AI-powered coaching platform, and a senior leader with responsibility for teaching, learning and coaching. This blog is part of a wider mission to support educators through meaningful reflection, not performance metrics. It documents the journey of building Starlight from the ground up, and explores how AI, when shaped with care, can reduce workload, surface insight, and help teachers think more deeply about their practice. Rooted in the belief that growth should be private, professional, and purposeful, The Insight Engine offers ideas and stories that put insight, not judgment, at the centre of development.

🔗 Connect with me on LinkedIn: https://www.linkedin.com/in/adam-sturdee-b0695b35a/