Designing Agency-Rich AI Disclosure to Reverse AI Penalties and Narrow Demographic Disparities

This research proposes replacing binary AI disclosure labels with a participatory, agency-rich design that includes provenance toggles showing AI-suggested versus human-edited text, opportunities for readers to propose edits or flag passages, and human-authored notes explaining AI use. The goal is to test whether such disclosures transform AI labels from mechanistic cues into competence signals, thereby reversing AI penalties, reducing demographic disparities in reader judgments, and aligning human and LLM raters.

References:

1. How Does the Disclosure of AI Assistance Affect the Perceptions of Writing?. Zhuoyan Li, Chen Liang, Jing Peng, Ming Yin (2024). Conference on Empirical Methods in Natural Language Processing.
2. Penalizing Transparency? How AI Disclosure and Author Demographics Shape Human and AI Judgments About Writing. Inyoung Cheong, Alicia Guo, Mina Lee, Zhehui Liao, Kowe Kadoma, Dongyoung Go, Joseph Chee Chang, Peter Henderson, Mor Naaman, Amy X. Zhang (2025). arXiv.org.