Reproduce “The Psychological Costs of Artificial Intelligence Adoption in Software Engineering” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Adam Alami, Elda Paja, Abhishek Tiwari
A qualitative case study of generative-AI adoption at a large Danish software company identifies five psychological costs that can accompany perceived productivity gains: uncertainty distress, accountability anxiety, intensified cognitive load, disruption of craft identity, and erosion of meaning and satisfaction. Interviews show that practitioners manage these pressures through staged delegation, additional verification, restricted AI use, and deliberate preservation of agency, competence, and professional identity.
Key insights
- The study identifies five recurring costs: uncertainty distress, accountability anxiety, cognitive-load intensification, craft-identity disruption, and erosion of meaning and satisfaction.
- AI can reduce code-production effort while creating a verification tax: extra work to inspect, test, explain, correct, and justify generated outputs.
- Psychological costs are not equivalent to resistance; participants often accepted AI’s benefits while limiting use because of governance constraints, unsuitable tasks, resource limitations, or concerns about agency and competence.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.