The best way I know to do this is to make a table and assume you have a population of 100000 people.
The table looks like this
test negative test positive
not infected 95952 3998
infected 2 48
How did this table get produced?
In 100000 people 50 are infected.
Since the test is positive 96% of the time when a person is infected, there will be 48 people in the lower right box and that leaves only 2 in the lower left box.
There will be 99950 people not infected. Of those 4% will have a positive test, i.e. 3998 and the other 95952 not infected will have a negative test.
Now your probability should be easy to see.
Part A: 48/(3998+48)
Part B: 95952/(95952+2)