People with the AB-positive blood type have all three of the A, B and RhD antigens present on their red blood cells, and no antibodies to these in their blood serum (aka blood plasma.)
This means that they cannot receive blood serum transfusions from people of other blood types, as their serum does contain these antibodies. In general, a person's blood serum contains antibodies for whichever of the A/B/RhD antigens are not on their blood cells, and no others.
Otherwise, an acute haemolytic reaction would occur.
The most-upvoted answer to another question on Biology.SE used a quote from Dean's "Blood groups and red cell antigens", chapter 5 to explain why these antibodies were present:
ABO antibodies in the serum are formed naturally. Their production is stimulated when the immune system encounters the "missing" ABO blood group antigens in foods or in micro-organisms. This happens at an early age because sugars that are identical to, or very similar to, the ABO blood group antigens are found throughout nature.
In a hypthetical future situation where the demand for blood serum transfusions was high, would it be possible to vaccinate young AB+ children against these antibodies? I'm thinking of injecting proteins similar to the antibodies, to train the immune system to produce "anti-antibodies". The memory-B cells would hopefully produce these fast enough and in large enough quantities to destroy the original antibodies and prevent ill-effects when a transfusion was given.
Would such a vaccine be feasible?
If not, would an immunisation relying on very small quantities of the actual antibodies (not enough to damage/agglutinate significant numbers of red cells) be viable?
I'm aware that if/when such antibodies pass through the placenta, they can lead to haemolytic disease of the newborn, so a newly-born child might be too young to receive such a vaccine. I'm imagining a situation where the number of such antibodies would not be high enough to cause these complications, or where the fake antibodies would not bond to the red cell antigens.
Dean, L. (2005). Blood groups and red cell antigens. 1st ed. Bethesda, Md.: NCBI.