Adaptive Comparative Judgement and AI generated content.

By Mark House

At the core of the RM Compare system is a machine learning algorithm - it's what delivers the adaptivity that makes the comparative process so powerful. The advent of new ways of providing AI generated content opens up exciting new possibilities, for example in Learning by Evaluating contexts.

What is Learning by Evaluating?

Learning by Evaluating (LbE) is a common use case for RM Compare. The short video (below) provides a good overview of research led by some of our partners in the US that showed how a simple comparative process resulted in dramatic learning gains for all students.

LbE brings assessment to the start of the learning process by allowing students to compare items that are similar to the work they are about to complete. In doing so we can help students to achieve 3 key things:

  1. Empower students to learn independently,
  2. Help them understand what good quality work looks like,
  3. Improve their own learning outcomes.

The LbE process is reliant on students having access to stimulating work to compare. To this point this has most commonly been work from previous cohorts.

Sourcing this material can sometimes be challenging and time consuming, however with new AI authoring products there is perhaps an alternative approach. Chat GPT-3 for example can be prompted to generate written content, while Dall-E-2 can create remarkable images from a simple description in natural language.

How can I try this?

If you are already an RM Compare license holder you can get going right away. You might for example launch a session consisting of only AI generated content for your students to compare. Alternatively, you might want to mix AI generated content with non-AI content. You could even ask the students to contribute directly to the session as part of the learning process.

If you don't currently have a license, you can get going right away by signing up for the free trial using the link below.

Is there anything else I could do with this type of content other than LbE?

Almost certainly! We can't wait to see how users incorporate AI generated content into RM Compare. Remember that there are already a large number of different item types that can be used in the system including images, documents, video, audio and webpages.

We appreciate that we are just at the very start of what is sure to be an unprecedented period of transformation.