From Idea to Implementation: How ChatGPT Streamlined Proficiency Scale Creation at Colegio Bolivar

By Jon Roberts, Middle School Principal and Katie Neal, Innovation and STEAM Coordinator

When ChatGPT was released in 2022, we all thought that big changes to education were imminent. One year later, at Colegio Bolivar, we discovered that not much had changed. Our data showed that students knew about generative AI but were mostly just experimenting with it and teachers were not using it or learning about it. We recognized AI’s potential by writing an initial position statement, but the ideas were not yet being implemented. That same year, all of our teachers had taken on the complex task of learning about and creating proficiency scales. Here’s how Colegio Bolivar used ChatGPT to help teachers learn about AI, reclaim time, and create high quality proficiency scales.  

Giving Teachers Time and Purpose to Learn AI

Entering the 2024-25 school year, we knew that only a fourth of our teachers had spent at least an hour learning about generative AI tools, and only about 10% were using them. We also knew that our continued focus on proficiency scale creation would require a lot of teacher time and brainpower. Time was becoming a barrier to implementation, and we knew we had to do something to streamline the process of creating proficiency scales so that we could focus on the conversations about assessment and planning for deeper learning.

A lot of our teachers weren’t jumping on ChatGPT to draft their proficiency scales right away—they were still figuring it out, unsure about it, or busy with other priorities. As an administrative and learning team, we decided to intentionally bring AI into our designated times for proficiency scale creation. This gave teachers the opportunity to receive guided instruction about ChatGPT while also meeting their goal of completing their proficiency scales. After seeing the speed and ease with which they were able to generate their first drafts of proficiency scales, teachers felt empowered to meet the goal of finishing all of their scales by December 2024. 

ChatGPT is Only Part of the Process

Our incorporation of AI intentionally included the human touch as well. We ensured that a real person was prompting, interpreting, and revising all of the aspects of generative AI at each step. To create the first drafts of proficiency scales quickly, our secondary instructional coach (and resident prompt engineer) worked with individual teachers to design and refine a prompt that all teachers could use.

The prompt was based on the proficiency scale expectations established by our Leadership PLC team and was refined based on individual teacher feedback on the proficiency scales generated by ChatGPT. The prompt was shared with all of our teachers along with a list of AI guidelines and questions to ask yourself about the ChatGPT-generated proficiency scale. We also scheduled collaboration sessions where departments would meet to share their scales and give each other feedback.

As of November, most teachers have completed all of their scales! With a clear goal (and a tool to achieve it) and accountability through the sharing sessions, there wasn’t much room left for resisting the process. The train was already moving. It was either get on, get support, or be left behind.

We see AI breaking down the initial barrier to getting started on challenging tasks. If we provide support and accountability, we believe it can transform practices and beliefs at a faster pace allowing more time to mentally digest the challenging work of teaching and deeper learning.