Beyond Detection: FARABI and the Assessment Credibility Shock in Higher Education

Generative AI is creating a credibility shock for higher education assessment. The issue is not only misconduct but a weakened evidentiary link between student work and the learning claims behind grades and credentials. Surveys and guidance from 2024 to 2025 indicate rapid AI uptake and that detection cannot provide assurance. This paper argues that traditional assessment systems, including take-home coursework graded mainly on the final product and invigilated exams, were already noisy evidence for higher-order learning. Generative AI accelerates these weaknesses by making plausible outputs cheap and by displacing tool influence into preparation and rehearsal. In response, the paper introduces FARABI (Framework for AI-Resilient Assessment and Balanced Integrity), a risk-based triage method for assessment portfolios. FARABI supports proportionate controls while acting as a diagnostic that shifts assessment toward applied judgment, observable process, and accountable tool use. The paper closes by questioning higher education’s warrant if employability weakens as the organizing metric.