[Piskvorrr]

Well, the problem with current state-of-the-art flight automation is that it can deal well with routine, repetitive tasks. What it fails at are corner cases (cf. US Airways Flight 1549), and self-diagnostics (sensor readings are inconsistent with current flight profile, e.g. Air France 447). In such cases, what you need is intelligence, not automatics.

This makes your argument another case of "let's assume we have strong AI; then pilotless aviation is easy-peasy." In other words, one of your unstated assumptions is still firmly in the realm of science fiction.