Karl Fogel at
This tale in today's @NYTimes is genuinely shocking. Grading in schools is problematic to begin with, but all defensible uses involve evaluating new information not remixing old information. IB and Ofqual were fundamentally confused about their task.
It's an excellent Op-Ed, but one quibble: the problem wasn't the use of "algorithms". All grading involves algorithms, even when it doesn't involve computers. The problem was the *inputs* to the algorithm. What was being "graded" was others' expectations about the student.