Plenty of writing advice says to sound like a college student. That instruction is meant well and it produces exactly the flat, transition-heavy register that detectors treat as machine writing, because the model is choosing the statistically safest way for a student to sound.
A person does not sound generic. They have a habit of starting with a caveat, a favourite way of connecting two ideas, one topic they will always over-explain. Those habits are what make prose recognisable, and they are also what make it read as human to a detector and to a reader.
So instead of asking what a student would sound like, ask what you would say first if a friend asked about the assignment. Write that sentence. Then write the second sentence as if you had to defend it.
There is a longer version of this argument on how to humanize AI text, and the free humanizer is a reasonable second opinion on a paragraph that still reads too even.