The proliferation of generative artificial intelligence challenges the credibility of assessment in higher education. This article advances a theoretical argument that universities must move beyond detection-based strategies towards ethically grounded, validity-driven assessment practices. Drawing on Ajzen's Theory of Planned Behaviour, Bandura's Self-Efficacy Theory, and situational crime prevention models, it analyses how AI exacerbates existing vulnerabilities within massified, commodified education systems. Technical countermeasures, including digital proctoring systems, are critically evaluated and found insufficient as standalone solutions. The case of Baird and Clare is used to illustrate how rehumanised, collaborative assessments can mitigate misconduct by enhancing student agency and ethical engagement. The article argues that safeguarding academic integrity in an AI-saturated era demands a fundamental pedagogical realignment, restoring the intrinsic purposes of higher education and resisting the instrumental rationalities that underpin surveillance-based governance.