Artificial Intelligence or Nursing Student? Revisiting Clues in the Connectives.

Abstract

Background: Recent research at a single-purpose nursing institution has suggested a means to authenticate student writing by distinguishing it from artificial intelligence (AI)-generated text through the detection of key terms.

Purpose: The purpose was to replicate and expand the research that identified key terms present in student writing but absent from AI-generated text.

Methods: A total of 5 generative AI writing tools were fed prompts to collect 14 787 words. Using the Search function on word processing software, the frequency of the terms, because, since, so, then, thing, think , and too , was measured and compared against earlier published findings from AI and students.

Results: The replication study was successful for the terms since, then, thing, think, and too.

Conclusions: Measuring key term frequency may be a path to authenticate student writing. While no tool can provide certainty of original authorship, the absence of key terms in a student submission may suggest AI authorship.