If $W$ is a finite Coxeter group, then it has a unique longest element $w_0$. In particular, $w_0$ is an involution.
This is a (paraphrased) quote from Fayers' paper $0$-Hecke Algebras of finite Coxeter groups. However, considering $S_3$, we have the longest elements are $(1 \; 2 \: 3)$ and $(1\:3\:2)$, neither of which is an involution and, of course, they're not unique. So how is the statement to be understood?