What Does the Credential Still Certify? Cognitive Stewardship for AI-Mediated Education

Credentials are losing their meaning without clear rules: When students use AI to complete assignments, universities can no longer reliably verify that a credential actually proves the student learned what it claims to certify. The paper argues this isn't just about catching cheating—it's about fundamentally rethinking what a degree proves.

Most university AI policies focus on the wrong thing: An audit of 30 universities' public AI guidance found that institutions are getting better at classifying whether students can use AI, but failing to explain what evidence still proves learning happened and how to protect the integrity of that evidence. Simply monitoring AI use isn't enough.

Universities need "cognitive stewardship" frameworks: Rather than just permitting or banning AI, the paper proposes institutions should clearly define four things: (1) what the learning claim is, (2) where the boundary is between student work and AI work, (3) what evidence proves learning occurred, and (4) what safeguards protect fair assessment.

Current policies are inconsistent and incomplete: The audit revealed that university guidance is clearest when AI is used as a final output substitute (like writing an essay), but becomes vague and uneven when AI is used for feedback, access, verification, or professional workflow—leaving gaps in how institutions actually protect credential validity.

Permission categories alone won't solve the problem: Universities need policies that make the "certification logic" transparent to students and employers—explicitly stating what cognitive work students can delegate to AI, what they must still demonstrate themselves, and how the institution will ensure fair evidence of learning.