Can AI mark controlled assessments?
As a teacher who's been experimenting with AI marking for BTEC controlled assessments, I have a few observations to share with you.
There is no doubt that marking a class set of controlled assessments can be a daunting task. We need to set aside a sizeable chunk of time, ensure that we have a quiet environment, and maintain our levels of concentration so that we get it right. I find the combination of time, quiet, and high concentration levels difficult to come by in a busy school environment.
While AI can't replace the human element entirely, it's certainly showing promise in assisting me with specific aspects of controlled assessment marking. Here’s what I’m finding:
What is the problem?
Sitting down to read through a set of BTEC written assessments can be a daunting task. You’ve taken the outside route (usually in the rain) to get to a quiet corner of the staffroom (essential because someone else is using your classroom) and you’ve laid out your laptop (after clearing the table of random clutter). You sit, determined to ‘get your marking done’ but you can’t start. After grabbing a coffee, you finally make a start but you find yourself reading the same paragraph over and over again. Does this sound familiar?
What have I done to help myself?
I’ve created a prototype for an app that will suggest marks and feedback for BTEC coursework. I can select the qualification and learning aim, paste in the work then hit the ‘Mark’ button. After a couple of seconds of thinking time, the app returns a suggestion for an appropriate mark band, mark and highlights the strengths and weaknesses of the work.
Does it work?
Yes, it works. The app points me to the best and worst bits of the work. I still have to read the work but the task is less demanding for me because a big part of the decision-making has been done. Because I’m confirming the mark and feedback rather than generating it myself, the cognitive load is lessened.
Are you cheating?
I don’t think so. I think I’m just working a bit more intelligently. I think that I’m more aware of what I’m reading because the decision on ‘what mark to give’ has already been made. But of course, that decision has not been made! I will decide once I’ve read the work, and not before. As I’m reading I will be comparing the feedback from the AI with the work in front of me. And what do human brains love when it comes to decision-making? Comparison! Websites like No More Marking offer a solution that uses Comparative Judgement for marking and it works beautifully!
But my little app prototype is simpler and requires less setup. As a tool for reducing cognitive load and soothing decision paralysis, it is wonderfully effective.
AI can help me to identify common errors like plagiarism, missing sections, or exceeding word limits, allowing me to focus on more nuanced aspects of the assessment. For tasks with rubrics, AI can generate basic feedback highlighting areas where students met the requirements for a particular marking band.
It speeds up the marking and makes it more accurate.
Conclusions
Overall, I view AI in BTEC-controlled assessment marking as a valuable tool to streamline certain tasks and provide basic feedback. Still, it's not a replacement for human expertise and judgment.
My role is to use AI effectively, alongside my experience and knowledge, to deliver fair, accurate, and meaningful feedback supporting student learning and development.
Health Warning
If you don’t know how to craft a prompt you will get inaccurate and unexpected results.
Transparency and communication with students about how AI is used in marking are essential to building trust and understanding.
Professional development for teachers on AI and its ethical implications is crucial for responsible and effective use.
Remember, AI is a tool, and like any tool, its effectiveness depends on how it's used. As I continue to explore and experiment, I believe AI has the potential to enhance, not replace, the human touch in many of our everyday teaching tasks ultimately benefiting our students.