GLOBAL A student submits an essay. You run it through three leading AI models. ChatGPT gives it a four out of five; Gemini agrees, but Claude insists on a perfect five. What score does the student actually deserve? Now, scale this scenario: what happens when one state-province uses ChatGPT to grade its high school graduating exams and another uses Claude? Are their GPAs, and the students who hold them, truly comparable? These are no longer theoretical questions.