I have one question about "Colossus:the Forbin Project" and all the other computers/robots gone wild movies. Aren't scientists smart enough to install off switches or atleast have a plug they can pull? Just remember: I was as good as any and better then most-Vincent Freeman in Gattaca
That would seem to be sensible. In the Forbin Project, however, I can only think of a couple possible scenarios that would make them choose not design a "kill switch" into the system: 1)Since they had tested Colossus prototypes in various scaled down forms, they had such confidence in their machine that a "kill switch" would be totally unnecessary. Or, 2) The existence of a "kill switch" would leave Colossus vulnerable to a very simple, single enemy act of sabotage: just pull the plug. Neither is an entirely satisfying explanation, but sufficient enough that it doesn't detract too much from the film.
Besides, the computer had already established a network with the twin computer in Moscow. The scientists would have to infiltrate and pull the plug on both machines simultaneously, not an easy job considering that the computers were effectively monitoring all of their vulnerable spots. At one point in the film the humans tried to regain control of the rocket silos and failed miserably, prompting a catastrophic retaliatory response.