I understand, but by working with the wrong units, the calculated force is 1 order of magnitude higher. By the way, even calculating with the wrong units, the answer should be 2x10^7 N instead of 2x10^8 N, which leads to a total error of two orders of magnitude higher.
I see that you are trying to estimate an average crushing force for a vehicle traveling at 90 km/h and coming to a sudden stop in 1 ft. This calculated average force will generate a deceleration of 100 g's, which is well above the maximum allowable deceleration of 40 g's and average of 20 g's on vehicle collisions, in order to protect the occupants. The crumple zone is designed to increase the duration of the collision, thus reducing the impulse. Impulse is Imp = mass x velocity and collision time will be coltime = Imp / Force.
Currently, the required frontal offset impact test is performed at 64 km/h (17.8 m/s) against a deformable barrier simulating a similar vehicle. Thus, considering the energy being dissipated over 1 ft (vehicle) + 1 ft (barrier) (0.6 m) will generate around 27 g's of average deceleration (17.8²/(2*0.6*9.81)).
So far, trying to answer the question "Are seats in rear crumple zones unsafe?", you can say no, if the impact energy is within the required design limits (offset impact at 64 km/h with a vehicle of same size), any other condition above this threshold will make increasingly unsafe to seat close to the crumple zones, but occupants seating on the other vehicle might also experience a deceleration above the allowable and might get severely injured even without being trapped by the collapsing structure, thus making airbags also highly desirable. There is also the probability factor guiding vehicle structural design, as most rear end collisions happen when both vehicles are going in the same direction with a difference in speed below 60 km/h.