A learning platform uses an AI to automatically select tasks for students.
It should give each student questions that match their current level.
At first, the system seems to work well.
Some students receive more difficult tasks, others easier ones.
But after a few weeks, teachers notice a problem:
The AI labels some questions as “easy” even though many students struggle with them.
Other questions are considered “hard” even though they are solved quickly.
The platform was supposed to make learning fairer.
Why does the AI misjudge the difficulty so often?
Question:
Why can a school AI incorrectly estimate how difficult a task really is?
Solution will be unlocked tomorrow.