Anticipatory cognitive governance: A missing taxonomy for sovereign AI

Abstract

To support the next phase of European AI governance, this paper asks what forms of knowing, judging, and acting already exist within public administration—and how these should shape the design of artificial intelligence tools introduced to work alongside public servants. The notation Alternative Intelligence [ai] is introduced here to name this prior field of institutional practice, and we propose a taxonomy of four types as a missing foundation for sovereign integration. Current regulatory approaches address algorithmic accountability but neglect the cognitive infrastructure required for meaningful oversight. A complementary framework of anticipatory cognitive governance is proposed to support existing regulation such as the EU AI Act, positioning cognitive autonomy as a condition for legitimate public service delivery. A Cognitive Transition Audit is outlined as a diagnostic concept to assess institutional readiness before deployment.