Beyond AI Autonomy: Accountability Gaps, Loss-of-Control Risk, and Responsible AI Adoption

The study draws on survey data collected from approximately 550 respondents across multiple countries to examine the relationships among perceived superintelligence capability, AI autonomy, accountability gaps, AI safety governance, loss-of-control risk, organizational resilience, and responsible AI adoption. Respondents represented different professional positions and organizational contexts, with participants reporting varying levels of familiarity with AI and involvement in organizational AI adoption decisions. The data were analysed using partial least squares structural equation modelling (PLS-SEM) in SmartPLS to assess the measurement model and examine the hypothesized relationships among the constructs. The analysis included assessments of indicator reliability, internal consistency, convergent validity, discriminant validity, explanatory power, effect sizes, and predictive relevance. The dataset provides a basis for examining how perceived AI capabilities, governance arrangements, accountability challenges, and organizational conditions relate to responsible AI adoption.