Here is my proof:
Since $f$ is integrable, $\overline{I}_a^b(f)$=$\underline{I}_a^b(f)$. However, it is also a fact that
$L_f(M) \leqslant \underline{I}_a^b(f) \leqslant \overline{I}_a^b(f) \leqslant U_f(M)$, meaning that there exists a $\epsilon>0$ such that $\overline{I}_a^b(f) \leqslant \underline{I}_a^b(f)+\epsilon$,
and therefore $0\leqslant \overline{I}_a^b(f)-\underline{I}_a^b(f)\leqslant \epsilon$. By the general inequality I pointed above, since the lower
integral is the supremum of lower sums $L_f(M)$ over all partitions and the upper integral is the
infimum of upper sums $U_f(M)$ over all partitions, $U_f(M)-L_f(M)<\epsilon, \ \ \forall \epsilon>0$
Could anyone correct my proof as I am struggling with the $\geqslant \gt$ signs?
Let's let $I$ be the integral of $f$, which exists and is finite by assumption. Because it is the sup of the lower sums, for any $\epsilon_1>0$ there are infinitely many partitions $M$ so that $I-L_f(M)<\epsilon_1$. Similarly, there are infinitely many partitions $N$ so that $U_f(N)-I<\epsilon_2$. If $\epsilon>0$, and we choose $M$ and $N$ with epsilons given by $\epsilon_1=\epsilon_2=\epsilon/2$, and let $P$ be the common refinement of $M$ and $N$, what can we say about $U_f(P)-L_f(P)$?