As universities expand AI use in student work, a growing number are removing AI-detection tools from assessment workflows and replacing them with new integrity and evaluation approaches. The argument from institutions: AI detectors are unreliable indicators of cheating. The policy shift puts faculty in the spotlight, because it changes what instructors must do to uphold academic integrity while also reflecting the reality that students increasingly have AI assistance available. The article describes how assessment redesign is becoming the alternative to “technological enforcement,” requiring changes to how work is produced, reviewed, and contextualized. The emerging governance challenge is operational and pedagogical: how faculty can reduce incentives for misconduct while avoiding policies that rely on flawed detection methods. This transition also raises questions about consistency across courses and the support faculty need to adopt alternative assessment designs.