Augmenting Critical Thought: Designing a Pedagogical Framework for Generative AI Based on Educational Perceptions
Generative AI tools such as ChatGPT and Gemini have become increasingly common in higher education. Students use these tools for a wide range of academic tasks, including brainstorming, summarizing texts, structuring assignments, and generating code.
While these technologies offer significant benefits in terms of efficiency and accessibility, concerns have emerged regarding their potential impact on critical thinking, independent problem-solving, and genuine learning. As AI becomes more integrated into educational environments, universities face the challenge of ensuring that students remain cognitively engaged rather than becoming overly dependent on automated systems.
Objective
The objective of this bachelor's thesis was to investigate how students and lecturers perceive the use of generative AI in higher education and to develop a pedagogical framework that supports meaningful AI-assisted learning. The research focused on three key questions:
- What distinguishes Cognitive Augmentation from Cognitive Offloading?
- Which academic activities are most vulnerable to passive AI reliance?
- What educational checkpoints can help maintain meaningful learning when AI is used?
Methodology
The study followed a three-phase research design. First, a literature review was conducted to examine existing research on artificial intelligence, critical thinking, cognitive offloading, and educational theory. Second, survey data was collected from students and lecturers at the Bern University of Applied Sciences (BFH). The survey combined quantitative and qualitative questions to capture both usage patterns and perceptions of generative AI. Finally, the findings from the literature review and survey were combined to develop a practical pedagogical framework for AI-supported learning environments.
Key Findings
The results showed that generative AI is already deeply integrated into student workflows and is regularly used for tasks such as brainstorming, assignment structuring, text summarization, and code generation. Several findings highlighted potential risks associated with excessive AI reliance:
- 54.1% of students reported lower confidence in solving problems independently.
- 50% stated that they sometimes struggle to complete academic tasks without AI assistance.
- More than one-third of students believed their critical thinking abilities had weakened due to AI use.
Lecturers reported similar concerns. While student submissions often demonstrated improved grammar and presentation quality, lecturers also observed reduced analytical depth, less originality, and increasingly similar writing styles.
The study further identified coding and text summarization as the academic activities most vulnerable to Cognitive Offloading, while discussions, workshops, and oral explanations were found to carry significantly lower risk.
Proposed Framework
Based on these findings, a pedagogical framework was developed to encourage Cognitive Augmentation while reducing Cognitive Offloading. The framework is built on four principles:
- Process over Product
- Rebuild Confidence in Independence
- Active Verification of AI Outputs
- Flexible and Accessible Assessment
To support implementation, the framework proposes reflective verification activities, light verbal checkpoints, and flipped evaluation methods where students critically assess AI-generated content rather than passively consuming it.
Conclusion
The findings suggest that generative AI is neither inherently beneficial nor inherently harmful. Its educational impact depends largely on how students use these tools and how universities design learning environments around them. Rather than focusing solely on whether students use AI, educational institutions should focus on ensuring that students continue to think, reflect, verify, and engage critically while using AI-supported technologies. The challenge for higher education is therefore not to resist artificial intelligence, but to integrate it in ways that preserve and strengthen meaningful human learning.