Motivation

Recent developments in Generative AI and Agentic AI have the potential to fundamentally transform knowledge workby providing powerful support for problem solving. While such AI-based assistants provide unprecedented access to information and streamline communication, a recent MIT study demonstrates deteriorating effects on users’ critical thinking skills, which implies the risk of intellectual deskilling. Users are at risk of merely consuming AI-generated answers without actively engaging in reasoning and reflection, thereby losing their own problem-solving and communication skills.

Goal

The overarching aim is to explore mechanisms that help users enhance their skills when solving problems with AI-based assistants. Based on preliminary works in the context of conventional AI, we assume that reflection mechanisms that induce active cognitive engagement in the interaction with an AI assistant (e.g., choosing between different options, intervention possibilities, explanations alongside AI-generated answers) can enhance critical thinking skills. The goal of the seminar is to develop such reflection mechanisms and investigate their impact on user behavior.

Exemplary research questions (students can specialize)

• How can mechanisms be designed that prevent intellectual deskilling while preserving efficiency gains in AI-assisted knowledge work?
• How do different reflection mechanisms influence users’ critical thinking when solving complex problems with AI assistants?
• How do autonomy and reflection influence the risk of intellectual deskilling in AI-assisted knowledge work?

Suggested work plan (students can specialize)

• Systematic literature review: Identify and analyze conceptual and empirical work on deteriorating effects of AI usage with a focus on deskilling.
• Theorizing: Develop a theoretical framework that explains how reflection mechanisms can prevent intellectual deskilling effects in AI-assisted tasks.
• Design and development: Design and iteratively develop reflection mechanisms for AI-assisted knowledge work, following the design science research paradigm.
• Empirical investigation: Conduct controlled experiments to investigate the effects of reflection mechanisms on users’ critical thinking skills.

Expected deliverables

• A short (12–15 pages) paper-style report
• A short pitch (5 min) and discussion of findings
• Optional: reproducible codebase (e.g., Python)

Contact:

Sebastian Schäfer, Sebastian.schaefer2@kit.edu
Information Systems IV – Digital Platforms and Services

Starting points

• https://time.com/7295195/ai-chatgpt-google-learning-school/
• Förster, M., Schröppel, P., Schwenke, C., Fink, L., & Klier, M. (2024). Choose Wisely: Leveraging Explainable AI to Support Reflective Decision-Making. International Conference on Information Systems.
• Webster, J., Watson, R. T. (2002). Analyzing the Past to Prepare for the Future: Writing a Literature Review. MIS Quarterly, 26 (2) xiii-xxiii.
• Peffers, K., Tuunanen, T., Rothenberger, M., Chatterjee, S. (2008). A Design Science Research Methodology for Information Systems Research. Journal of Management Information Systems, 24 (3) 45–78.