This line of thinking only really makes sense if you presuppose that every person taking a computer science course in high school has an understanding of file types. Many don't have this knowledge. I don't see why students should need to have additional knowledge that is outside the scope of a course, just to pass the clase.
There's a reason these things are taught in a CS program, often at higher levels. And that's because they're not common knowledge.
Besides, I don’t know what it’s like in the US nowadays, but back in my time (late 1990s, IB) the computer science curriculum took file formats and so forth as granted, and concentrated on the theoretical constructs such as Turing machines, lambda calculus, decidability, computability, and the theory of recursive functions. I remember my exam required proving the identity between a particular integral and the Euler (not Riemann) zeta function and whether it could be evaluated in polynomial time to an arbitrary degree of approximation. We had to sketch out the algorithm that would do this in terms of pseudocode.
Trying to transcode a file by changing the extension... I dunno... it seems kind of meek.