I started with the inclusion exclusion principle. Let
$B_{1} = A_{1}, B_{2}=A_{2} - A_{1}, \ldots , B_{n} = A_{n} - (\cup_{i=1}^{n-1}A_{i})$
So, $A_{n} = B_{n} \cup (\cap_{i=1}^{n-1}A_{i})$.
We know that $P(\cup A_{i}) = P(\cup B_{i}) = \sum P(B_{i}) = \sum P(A_{i}).$
Also, $P(A_{n}) \leq P(B_{n}) + P(\cap_{i=1}^{n-1}A_{i})$ and $P(A_{n}) \geq P(B_{n})$.
But I am not sure what to do after this. Also, the question asks for countable events and I am doing this for finite events. Any help will be appreciated.
Ιf existed $A_j,A_i$ such that $P(A_i\cap A_j)>0$ then
$P(A_i \cup A_j)=P(A_i)+P(A_j)-P(A_j\cap A_i)<P(A_i)+P(A_j)$
So $$P(\bigcup_kA_k) < P(\bigcup_{k\neq i,j}A_k )+P(A_i)+P(A_j) \leq \sum_{k \neq i,j}p(A_k) +P(A_j)+P(A_i)=\sum_kP(A_k)$$
which is a contradiction