Bloomy: Measuring Cognitive Development and Detecting Unproductive Cognitive Offloading in Workplace AI Interactions
Authors/Creators
Description
Generative AI has broken a long-standing assumption of workplace learning: that the quality of an employee's output reflects the quality of their thinking. When an assistant can produce a polished plan, summary or critique on request, organisations can no longer see which thinking their people still do themselves.
This concept paper introduces Bloomy, a framework that works within employees' everyday AI interactions. It measures how each employee's own thinking develops over time across the 19 cognitive processes of the revised Bloom's taxonomy, and detects when AI use shifts from saving time to eroding skill. Bloomy operates inside a single chat, in Arabic, Franco-Arabic or English, through short, optional invitations to think, with no exams and no surveillance. Its core principle is "think first, then ask": when a request contains the employee's own thinking, that thinking counts as evidence of capability.
The paper outlines the problem, the framework's design, its ethical commitments and a planned pilot with organisations in Egypt, Saudi Arabia and the UAE. The full specification is maintained privately.