Mitigating Overreliance on GenAI in Software Development

  • Type:Bachelor/Master Thesis
  • Date:As soon as possible
  • Supervisor:

    Tom Weinhardt

  • Add on:

    ARAI

Background and Motivation

Generative AI tools like GitHub Copilot or Cursor are rapidly transforming how developers write and review code. While these tools boost productivity, they risk fostering overreliance: developers increasingly accept AI-generated outputs without critical reflection. Drawing on dual-process theory, Cognitive Forcing Functions (CFFs) have been proposed as design interventions that promote reflective reasoning. Yet CFFs remain largely underexplored in the context of generative AI and software development, even though developers are among the groups most affected by GenAI adoption.

Objectives

This thesis investigates how CFFs can be adapted to the context of GenAI-assisted software development and evaluates their potential to reduce overreliance among professional developers in an online experiment. The exact research approach and scope can be defined collaboratively.

Profile

  • Interest in human-AI interaction and experimental design
  • Programming skills in Python
  • Basic knowledge of statistics and quantitative data analysis
  • Self-driven, structured, and detail-oriented working style
  • Good English skills

Contact

We offer a timely and relevant research topic with close supervision and the possibility of co-authorship. Please send your transcript of records, a short CV, and a brief motivation (max. ½ page) to Tom Weinhardt (tom.weinhardt∂kit.edu).

Starting Literature

Buçinca, Z., Malaya, M. and Gajos, K. (2021), “To Trust or to Think: Cognitive Forcing Functions Can Reduce Overreliance on AI in AI-assisted Decision-making”, Proceedings of the ACM on Human-Computer Interaction, Vol. 5 No. CSCW1, pp. 1–21, doi: 10.1145/3449287.

Evans, J. and Stanovich, K. (2013), “Dual-Process Theories of Higher Cognition: Advancing the Debate”, Perspectives on Psychological Science, Vol. 8 No. 3, pp. 223–241, doi: 10.1177/1745691612460685.

Zacharias, J., Popova, A., von Zahn, M., Chen, J. and Hinz, O. (2026), “Developers’ Dilemma: Opportunities and Pitfalls of Generative AI for Software Development”, Business & Information Systems Engineering, doi: 10.1007/s12599-026-00998-y.