Research Summary: The disclosure gap: why mandating AI transparency without protecting honesty is a leadership failure

Despite widespread mandates for generative AI disclosure by academic journals and universities, actual reported AI usage in published papers remains extremely low, creating a significant 'disclosure gap'. This discrepancy is not attributed to a failure in compliance but rather to institutions asking individuals to volunteer information, indicating a potential systemic issue in current transparency approaches.

Why it matters

The persistent gap between mandatory AI disclosure policies and actual reported usage highlights a fundamental challenge in implementing transparency measures for emerging technologies. This indicates that current institutional strategies may be ineffective in fostering genuine accountability and understanding of AI's integration into critical processes, necessitating a re-evaluation of governance and leadership approaches.

Key insights

• Approximately 70% of academic journals now require disclosure of generative AI use.
• Only about 0.1% of papers published since 2023 declare AI use, despite mandatory disclosure policies.
• Journals with disclosure policies do not show lower rates of AI adoption compared to those without such policies.
• Submission data from major publishers indicate disclosure rates of 3.3% to 5.7%, significantly lower than survey data suggesting 28% to 76% of researchers use these tools.
• The 'disclosure gap' is framed not as a compliance failure requiring firmer rules, but as an inherent issue when institutions rely on voluntary self-reporting.

Source

Frontiers in Education — https://www.frontiersin.org/articles/10.3389/feduc.2026.1966879