The generative AI paradox: GenAI and trust erosion, corrosion of information verification, and the death of truth

Generative AI can generate realistic content at scale at extremely low marginal cost, and the paper argues that the greatest risk is not isolated deepfakes but the gradual erosion of shared cognitive foundations and institutional verification practices. The authors use a layered framework of content, identity, interaction, and institutions, combined with cases of fraud, elections, and others from 2023-2025, to sort out mechanisms such as cost collapse, customization, and provenance gaps. The mitigation framework emphasizes that provenance infrastructure, platform governance, institutional processes, and public resilience are complementary, and warns that society may discount digital evidence as a whole.