Nevertheless, the tendency of the Phillips screw to cam out easily was later found to be advantageous when used with early power tools, which had relatively unreliable torque-limiting clutches. In this context, cam-out helped protect the screw, threads, and driving bit from damage caused by excessive torque.[4]: 85–86 A follow-up patent in 1942 further refining the Phillips screw design describes this feature and argues that if screw-driving clutches were perfect, a screw recess with zero vertical contact angles (and thus no axial cam-out force) could be utilized. However, it noted that such designs had proven unsatisfactory on assembly lines, as the driving bits would not disengage in time to prevent damage.[5
You’re right it wasn’t originally designed for it but later patents and iterations on the Phillips head driver specifically designed it in which is where I think the confusion comes from. I had always heard it was originally designed that way but I guess not.
AFAIK, this is a myth. The original patents don’t include the tapering, and someone else in the comments quoted the relevant parts from Wikipedia.
Cam out source
You’re right it wasn’t originally designed for it but later patents and iterations on the Phillips head driver specifically designed it in which is where I think the confusion comes from. I had always heard it was originally designed that way but I guess not.