I would be cautious assuming any wrongdoing. The hanky-panky here appears to be mostly the fault of the entire appraisal system, more or less supposedly managed by TAF/AQB, a bunch of idiots. Yes, by law this and that. But writing a test that requires say 70% of all questions to be answered correctly to provide an acceptable rate of confidence is not easy. I mean, exactly 70%, regardless of the questions? Why 70%? I have taken some tough graduate math courses where the highest score by the brightest was below 70% and the passing score was maybe 35% - because the professors are looking for the one out of a million - and also want to show what high "standards" they have.
What makes sense, and is common practice in the US, is to use a curve for sloppily written tests: Just flunk the lowest 40-50% or whatever. In the US, this is called "standardized" scoring, but is better called "relative" scoring.
Why should states have different standards? Of course, they have different educational standards. New Mexico, Mississippi and Louisiana have the lowest standards for high school graduates, Massachusetts the highest.
===
Now, as an example of fixed standards: Germany is or at least used to be a country where all Abitur (high school) graduates have to pass the same countrywide exam to graduate and go on to college. The Germans never grade on a curve. Their educators spend a lot of time designing a set of tests that assign a certain number of points to different questions.
Grok-3: "Each subject in the Abitur is assessed based on exams and coursework, with grades typically ranging from 0 to 15 points (where 15 is the best). These points are then converted into a final grade on a scale of 1.0 (excellent) to 4.0 (passing), with anything below 4.0 being a fail. The specific criteria for assigning points are determined by the state's educational authorities and are intended to reflect absolute performance standards rather than a relative ranking.
That said, some states or schools may apply moderation or statistical adjustments after the exams to ensure consistency across years or to account for unusually difficult tests. This isn’t a traditional bell curve where a set percentage of students must fail, but rather a calibration to maintain fairness and comparability. For example, if an exam is deemed exceptionally hard, the point thresholds for certain grades might be adjusted slightly, though this is not a universal practice and depends on the state’s regulations.
In short, the Abitur is primarily standards-based, not curved, but local practices can introduce some flexibility that might feel curve-like to students. If you’re curious about a specific state’s approach, let me know, and I can dig deeper!"
===
Well, do you really expect the appraisers at TAF or the AI are competent to create such tests? Of course not and in any case no one is going to pay for that. They just decide to flunk a certain percentage with the lowest scores. That's easy-peasy.
So, I say, you really have to blame TAF, probably the AI as well, for this lackluster testing system for its inadequacies.
And this gal, if what I read in your post is correct, - she is most likely just out for revenge or an "idiot" or both, soaking up your sympathy. But, in the end, don't be so sure the AI won't be exonerated for this situation that really has to be laid.first and foremost at the feet of TAF.
As far as CS, riding on his motorcycle, I am sure he couldn't care less.
(One fix for this problem would be for the AI to just report the percentage correct, rather than a pass/fail.)