Collaborative knowledge construction with generative AI: Exploring argumentative co-writing processes through n-gram and cluster analysis

Since the beginning of CSCL research, collaborative writing has been playing a pivotal role asa tool for learning and knowledge construction. In the study presented here, we ask to whatextent large language models may not only assist individuals in their writing processes butalso serve as a collaboration partner. For this purpose, we analyzed the writing process ofindividuals supported by ChatGPT. We introduce the use of recurring n-grams as a means fortextual uptake, that is, the extent and granularity with which human writers adopt and adaptAI-generated text. Based on the overlaps between the ChatGPT output and participants’ finaltexts, we identified clusters of text reproducers, integrators, and reconstructors. Participants inthese clusters differed not only in their subjective contributions and authorship, but also intheir prior use of ChatGPT and their affinity of technology interaction. Referring to theconceptualization of interindividual interactions as uptake events, we suggest that n-grams areadequate means to analyze the uptake process in AI-supported human writing. Our findingsshow that AI-supported writing comprises distinct uptake patterns that differ systematically inthe degree of textual reuse and perceived authorship, thereby revealing varying modes ofengagement in human-AI co-writing, ranging from passive uptake of AI-generated text tomore active and integrative forms of collaboration.