I like this article because it carefully steps through the question it addresses. Readers of this newsletter will know that AI-scoring has been available for a number of years now, and easily predates the recent generative AI boom. As Nick Potkalitsky notes, "Ohio uses discriminative AI. Its job is to classify and score existing text. You give it an essay, it returns a number: 1, 2, 3, or 4 points." It is trained on human-graded essays and after training does nothing but classify essays into separate categories. By contrast, "The AI teachers worry about, tools like ChatGPT, is generative AI. Its job is to create new text.? It's completely different, shouldn't be used for grading, and probably wouldn't be very good at it (by contrast, discriminative AI is often fairer and more consistent than human graders).
Today: Total: [Direct link] [Share]
