Worthy of trust? AI governance and the role of (dis)trust

In this article, we critically examine prevailing narratives to trust in artificial intelligence that prioritize trust as an unqualified good, arguing instead for a more balanced approach that implies a shift from “trust-maximization” to calibrated trust, grounded in citizen vulnerability and public accountability. Such critical appraisal is even more necessary as artificial intelligence systems become deeply embedded in governance, raising crucial questions about public authority and democratic legitimacy. Three core concerns are highlighted: the decoupling of trust from trustworthiness and the corresponding underestimation of the role of distrust, the neglect of vulnerability as foundational to trust, and the emergence of “trust externalities,” where AI adoption shifts costs of failure to citizens without recourse or choice. Emphasizing the value of both trust and distrust, we discuss the merit of a “trust but verify” paradigm, underscoring the dangers of manufactured trust. We also call for contextual, experience-based methods to study trust and stress the critical role of accountability to realign technological reliance with institutional responsibility and citizen protection in AI-infused public governance. We propose that accountability stands as an important linchpin to recouple the link between trust and vulnerability, enabling trust to play its regulating role against the technology’s abuse, misuse, or overuse.