Raw LLM Responses

Article isn't working for me. As a technology instructor and administrator of all the online testing in my school, I go to a lot of these trainings. For NJ, we used the PARCC (now the NJSLA). Last year in one of our meetings, it was discussed the use of AI to grade essays. It's absolutely terrible. You could get an "A" if you just structure the essay to hit all the "key terms" the AI is looking for. Simply "restating the question" will automatically bump you up a ton of points since it hits the keywords. We were told to have students basically omit ALL pronouns from future essays to increase their score and how having more sentences (even if they are nonsense) will increase your score purely due to character count. It's a game. They are applying a mathematical AI formula to grade subjective content. It's going to be unfair and inaccurate...but they (companies like Pearson) will save A LOT of money from training/paying people to grade them by hand. If we can't even put in a platform to assess student work/standardized tests....why even test them in the first place? It's bad data. It doesn't grade for content, nor can it. It can't contextualize the essay or make inferences.