<p>Many students now use Generative Artificial Intelligence (GenAI) in their coursework, but their willingness to disclose that use remains low and risky. Based on postdigital approaches to agency, the article considers student GenAI use disclosure in higher education. We argue that non-disclosure should not be treated as a failure of self-regulation, metacognitive monitoring, or violations of academic integrity. These explanations position student GenAI use inside the learner and overlook the educational ecology where disclosure becomes meaningful. Drawing from self-regulated learning, learning analytics, postdigital scholarship, and the first author’s office hours narrative, we shift our unit of analysis from student disclosure profiles to relations among learners, instructors, tools, assessment, institutional policies, detection, and cultural assumptions about language and intelligence. We argue that concealment of GenAI can function as a situated form of regulation in response to environments where honesty may carry academic, relational, or symbolic costs. We propose postdigital agency as a way to study GenAI disclosure as an ecological, relational, and ethical practice grounded in trust, care, and negotiated responsibility.</p>