This article doesn't get into the reason why Phillips and Robertson are tapered and prone to camming out: they were designed that way to prevent too much torque from being applied when tightening by hand.
Neither parts of the statement is true for Philips: the Philips screws were borne out of a better compatibility with powered tools, and while the cam out properties have long been claimed as a torque limiter (generally in the context of assembling aluminum aircrafts before powered tools had good torque limiting), that was not part of the original 1932 patent. A later 1949 patent / refinement does make claims on that front, but those are mostly in terms of limiting damage to the screw IIRC.
