I would have understood if Mathematica had given this answer
$$\text{the would-be limit is} \ e^{i0}=1 \ \text{and/or} \ e^{i\pi}=-1$$
which are the 2 so-called adherence points to the sequence, not the limit in fact because strictly speaking the limit doesn't exist.
Indeed the content of the parenthesis when $n \to \infty$ is equivalent to $\frac{-n^2}{n^2}=-1$, therefore our sequence is equivalent to $(-1)^n$, somewhat "constantly hesitating" between $-1$ and $+1$...
The conclusion : you are in GIGO context (Garbage In, Garbage Out): when faced to a nonsense situation the best software (?) on Earth can give you nonsense answers...
E^(2 I Interval[{0, \[Pi]}]), which I think should be interpreted as "all points approach the unit circle in the complex plane." But that doesn't make a lot of sense given that the sequence is real. – Michael Seifert Nov 17 '21 at 21:02