Without any “inside” knowledge, then a simple empirical determination would be that air pressure is still used for brake application because it is deemed still to be the most efficient and effective method, or if not, the benefits of a potentially better method are not sufficient to justify the costs of conversion.
Note though that spring applied, air released parking brakes have been used in railroad service since around the late 1960s, firstly I think on MU passenger cars, but later on some locomotives. This was essentially a carryover from on-highway heavy truck and bus practice, where spring applied, air released brakes had been used for emergency, secondary and parking purposes since the second half of the 1950s, although probably not widespread outside of North American until the second half of the 1960s. (See for example this US patent
https://patents.google.com/patent/US275 ... =US2754805.)
The fact that in both railroad and on-highway applications, spring applied brakes have always been used in conjunction with air applied primary braking systems creates the inference that spring application is not viewed as being satisfactory for service brake purposes.
With a spring applied brake, a potential disadvantage is that the force available declines as the spring extends, so that the longer actuator travel occasioned by brake block wear would result in reduced braking force. With air pressure application, the air pressure remains essentially the same throughout the piston stroke. Well, that is true for self-lapping straight air brakes and UIC-type graduable release automatic air brakes, but not quite for direct release automatic air brakes, where the ratio between auxiliary reservoir and brake cylinder volumes is one determinant of application pressure.
Cheers,