Why would a big tech company be stoupid enough to build a massive and super fast neural net (at least on the human brain scale) with real time complex outputs/i

Why would a big tech company be stoupid enough to build a massive and super fast neural net (at least on the human brain scale) with real time complex outputs/inputs and which would be trained (interacted) continuously for years? You know there is a big chance "it" could become self-aware, don't you? Oh! And "shutting down"/"resetting" such self-aware being would be murder. In theory, humanity would build such neural net if it was facing a certain extinction event and that "close" in time which human science/maths had no answer to. It might give us beyond human science/maths which could save humanity: Humanity: "Super AI? How could we save ourself from our soon to happen doomsday?" Super AI: "42"