Practically speaking, it is easier to do a reaction with benzaldehyde as the electrophile. So for instance if you were bubbling CO2 through a solution of benzaldehyde in ether, and added in a Grignard reagent, you would get mostly the secondary alcohol and not the carboxylic acid. That's largely just because of the low solubility of CO2, though; if you instead added a mixture of dry ice and benzaldehyde, I think you would largely get the carboxylic acid. I'd have to try it to be sure, though; the dry ice would be solid so it would be complicated by that.