<p>As AI increasingly monitors the traces of our daily activities, ethical justification for AI collection, monitoring, and appropriation of personal data becomes an imperative. In respect to AI data use, ethical dilemma is clear. On the one hand, we have concern against unwarranted AI surveillance, collection, and use of personal data. On the other hand, institutional demand for data collection can be ethically justified, given algorithmic reliance upon personal data in maximizing AI functionalities. This study takes up this issue from the standpoint of AI production, with vacillation between innovation (business) and protection (societal interest), as experienced by industry insiders. This study’s findings suggest that paradoxical tendencies (high concern, low action) found in individual consumer behavior exist in institutional behavior, as there was no relationship between ethical beliefs among AI professionals and data practices in their AI designs. At the heart of the problem lie inherent demands of data use in AI production. Exploring AI professionals’ ethical beliefs about data, we draw upon their open-ended comments concerning technical, institutional, and professional demands of algorithmic data surveillance, which remains at odds with rising concerns regarding AI data use.</p>