Assessment in the Age of Generative Artificial Intelligence

Abstract

The history of education is replete with examples of disruptive technology. The internet, computers, and word processors are just a few examples that have changed the way we learn and teach. While such technologies have been transformative, their affordances have largely been limited to the storage, retrieval, and processing of information. The recent proliferation of generative artificial intelligence like ChatGPT, however, heralds a new age—one in which technology can create information with human-like fluency.

For decades, researchers have argued that tools change how we think, act, and learn. Despite this work, pedagogical practices such as assessment tend to separate learners from technology in misplaced efforts to measure their independent abilities. Consequently, present assessment designs are unable to account for the interdependence between humans and tools that often characterises authentic practice and learning, let alone the capabilities of a tool that can create human-like art, computer code, and text.

In this talk, I describe the ways in which tools can affect cognition and action and outline an evidence-centred design approach for the assessment of learning in contexts where humans collaborate with generative artificial intelligence.