Note that since $g$ has an $L$-Lipschitz gradient, we have that $\frac{L}{2}x^Tx-g(x)$ is convex. (Check this by showing that the gradient of $h(x)=\frac{L}{2}x^Tx-g(x)$ is monotone, i.e. $\langle h(x)-h(y),x-y\rangle \geq 0 \ \forall x,y$). In particular, since $\nabla (-g)=-\nabla g$ which is also Lipschitz, $\frac{L}{2}x^Tx+g(x)$ is also convex. Thus
$$
f(x)+g(x)-\frac{\alpha-L}{2}x^Tx=f(x)-\frac{\alpha}{2}x^Tx+\frac{L}{2}x^Tx+g(x)
$$
is a sum of convex functions and therefore convex, so $f+g$ is indeed $\alpha-L$ strongly convex.
Similarly, we have
$$
g(x)-f(x)+\frac{\alpha-L}{2}x^Tx=-(\frac{L}{2}x^Tx-g(x)+f(x)-\frac{\alpha}{2}x^Tx)
$$
We know that $\frac{L}{2}x^Tx-g(x)+f(x)-\frac{\alpha}{2}x^Tx$ is convex as a sum of convex functions, so that $-(\frac{L}{2}x^Tx-g(x)+f(x)-\frac{\alpha}{2}x^Tx)$ is concave, and hence $g-f$ is $\alpha-L$ strongly concave.