I spent several years working on sustainability in international education before co-founding AskEd. So, when institutions ask me about the environmental impact of AI, I do take the question seriously.
The concern is understandable. AI has a real carbon footprint, and the sector is right to think carefully about what it adopts... and why. But I think the question is often framed in a way that leads to the wrong conclusion.
The debate tends to focus on AI in aggregate, which makes it look alarming. Data centres, energy demand, water consumption. The numbers at scale are significant and worth scrutinising. But that is not the relevant comparison for an institution deciding whether to use a conversational AI tool to handle student enquiries.
Google published research last August showing that a single text prompt on its Gemini platform uses 0.24 watt-hours of energy. That is roughly the equivalent of watching television for nine seconds. A typical AI conversation, even a longer one, sits in that range. For context, sending an email with a large image attachment uses more.
The comparison that matters for student recruitment is not AI versus nothing, it is AI versus the alternative: staff handling the same volume of enquiries by phone, email and travel, across time zones, in multiple languages, at all hours. That alternative has a carbon cost too, and it is considerably less scalable.
There is also a question about what the AI is being used for. Not all uses are equal. Generating images, producing video, running complex models for trivial purposes. Those are legitimate targets for scrutiny. Helping a prospective student in Brazil get an answer about course fees at eleven o'clock at night, in Portuguese, without anyone getting on a plane or picking up a phone, is a different category of use entirely.
International students are going to travel. That is not something AI changes. But the argument that facilitating international education is part of a virtuous circle is one I find persuasive. Students who study abroad tend to bring knowledge, skills and perspectives back to their home countries. Some will be involved directly in sustainability research. Most will carry a broader understanding of global challenges into their working lives. Helping them find the right institution, in the right way, at the right moment, is not a frivolous use of energy.
None of this means institutions should adopt AI uncritically. Measuring impact, understanding what you are deploying and why, and being honest about the trade-offs is exactly what good practice looks like. At AskEd we have a sustainability policy precisely because we think that accountability matters.
But declining to use a tool that could help a student get a timely answer to a question, on the grounds that AI is bad for the environment… is not really a sustainability position, is it? It’s more like the avoidance of one.
By John Crick
