Written by Gary Cossins, The AI Team, The English College, Prague

With the best will in the world, it’s difficult to break from routines and practices that have been embedded through years of teaching. One particular ritual that I have enacted is my twice monthly habit of writing to students imploring them to write more cohesively. Sometimes I vary my practice - I might dedicate a starter activity to the skill, or post an example to the classroom page, but always the same message ad nauseam with little effective change. It was after one cycle of this pedagogical dance that a student asked me what a more cohesive response might look like, and, as they caught me in a whimsical mood, I said they should ask AI. Within five minutes of collaborative talk, we had broken our cycle and created a new classroom routine. Students would input a sample of their writing, request that the AI re-write it with a specific goal from my feedback, and then reflect on what had been changed and why. Written responses improved, student’s retained knowledge, and this happened without me having to change my feedback.

The AI Team

As a zealous AI convert/apologist, I was invited to join my school’s AI Committee (a small working group of teachers with a shared interest). The group came about in recognition of the potential threat to academic honesty, but also the potential application of this emergent technology. Our leadership team recognised the need for a clear school policy on this, and tasked us with the creation of this document. After some research, we learned that nobody really had a working policy for AI use and that we were teetering on the vanguard of educational practice. We knew that we had to get ahead of the game - students were already using AI, teachers were starting to test the water.

Weapons of choice

Short of an extinction-level event, AI wasn’t going anywhere, so we quickly realised that a ban on the use of AI was neither practical nor desirable. Instead, we wanted to control the use of AI. With this aim in mind, we opted to licence Google’s Gemini for use across the whole school. We opted to extend our Google licence to include Gemini as they were the devil we already knew.

We found that there was a difference of perception between students and teachers. Many teachers were worried that students would use AI to avoid work, but students saw little difference between existing tools such as grammarly and spell check. Neither idea is without merit.

Next Steps

What can we do to ensure that students and staff use the tool responsibly? Firstly, we needed to remove the element of fear. Our exam route requires that teacher’s authenticate and submit coursework, so the notion of proving dishonesty was taken away. If a piece of work was suspicious, we would get a second opinion from within our faculties and if needed a conference would be held where the student would be asked to articulate their arguments verbally. We found the various AI checking software to be inconsistent, and led to direct conflict with the student. If a student is unable to persuade us of the authenticity of their work, we would give them a chance to repeat the task in controlled conditions. This helps to avoid the initial point of conflict, but it’s not a long-term solution. We know that the next step is in modelling responsible AI used to showcase creative thinking, and I look forward to seeing how our teaching staff can rise to this challenge.

Find out more about COBIS Training Schools