Let p(Ad) denote the probability that item drawn from A is defective, and

P(Bd) the corresponding probbaility for the item drawn from B.

Then Bayes theorem tells us:

p(Ad| Ad or Bd and not(Ad and Bd))= p(Ad or Bd and not(Ad and Bd)|Ad)p(Ad)/ p(Ad or Bd and not(Ad and Bd))

Now:

p(Ad)=3/8,

p(Ad or Bd and not(Ad and Bd))=3/8+2/5-(3/8)(2/5),

p(Ad or Bd and not(Ad and Bd)|Ad)=1-(2/5)

RonL