Implementing a proposed framework for enhancing critical thinking skills in synthesizing AI-generated texts

Abstract

Since the development of open and low-cost generative artificial intelligence (GenAI), the higher education community has witnessed high use of AI-generated texts in scholarly research and academic assignments, attracting ongoing debate about whether such practices constitute cheating. While scholars argue that integrating GenAI tools can enhance productivity, critics raise concerns about the negative effect of such integration on critical thinking (CrT). This study therefore proposed a framework for enhancing students’ CrT skills in synthesizing AI-generated information. The proposed framework is underpinned by various theoretical foundations, encompassing five interconnected step-wise phases (familiarizing, conceptualizing, inquiring, evaluating, and synthesizing). The study was conducted under two separate experiments. The first experiment (Study 1) validated the effectiveness of the proposed framework, providing CrT training to 179 postgraduate students. In the second study (n = 125), additional experiments were undertaken to confirm the effectiveness of the framework in different contexts. An experimental procedure involving pretest and posttest design was implemented wherein participants were randomly allocated to one of three groups: experimental group 1 (exposed to our framework), experimental group 2 (exposed to an alternative self-regulated learning framework), and a control group (exposed to a non-structured framework). Results from Study 1 revealed that the framework enhances students’ CrT skills to synthesize AI-generated texts. However, these CrT skills manifested through various rigorous training aimed at reinforcing learning. While the proposed framework holds considerable value in cultivating CrT skills, significant differences arise across various personality traits. In Study 2, the framework proved to be effective in different contexts. However, it did not make a difference, particularly in its capacity to enhance students’ self-regulated learning compared to other frameworks. We discussed the implications of the findings and recommended it to educators seeking to prepare students for the challenges of the AI-driven knowledge economy.