Developing Thinking through LLM-Assisted Writing: Hegelian Synthesis and Critical Thinking
Abstract
Students can bypass much of the writing process and the critical thinking that comes with it when using Large Language
Models (LLMs) such as ChatGPT. Single-stage writing assignments may have no value for students who use LLMs. This paper
proposes Hegelian synthesis writing (dialectic writing) as a solution for this problem. Dialectic writing requires
students to develop arguments in stages over time. The stages deepen perspective, lead to discovery, and may produce
original conclusions composed of conflicting viewpoints. While students can use ChatGPT to brainstorm and practice
thesis, antithesis, and synthesis essay form, this study shows ChatGPT does not evaluate texts truthfully and often
fails to produce strong thesis/synthesis statements. Instructors who want to promote critical thinking must have
students critically evaluate and revise ChatGPT outputs. Survey results from classwork using ChatGPT to produce
synthesis essays show students are receptive to using ChatGPT to brainstorm and learn essay structure. The results also
suggest students need more support to identify ChatGPT deficiencies in creativity, particularly with synthesis
conclusions. LLMs can model dialectic writing, but students need clear expectations for their role in the writing
process. In the age of LLMs, we must look to synthesize student and AI writing and have students emerge as better
thinkers. Assignments that require students to evaluate and revise ChatGPT outputs and to create new conclusions appear
best suited to produce this outcome.