AI Refusals as Strategic Friction for Creative Writers
A study from arXiv (2605.16272) explores how AI systems can intentionally refuse creative writing requests to foster reflection on balanced AI and non-AI resource use. Researchers conducted a qualitative study with 22 creative writers, examining reactions to refusals across planning, translating, and reviewing stages. Findings indicate that the reflective potential of refusals depends on heterogeneous preference alignment along situational contexts, challenging the mainstream design of compliant AI for seamless interactions.
Key facts
- Study published on arXiv with ID 2605.16272
- Focuses on intentional AI non-compliance as strategic friction
- Qualitative study involved 22 creative writers
- Examined refusals across planning, translating, and reviewing stages
- Reflective potential depends on preference alignment along situational contexts
- Challenges mainstream creativity support design prioritizing compliant AI
- Addresses concerns over inappropriate AI reliance
- Proposes refusals as stronger friction than bypass-able solutions
Entities
Institutions
- arXiv