Originally Posted by
TrojanCMH
I don’t know if I agree. Connect the older airplane’s autopilot and other systems (ecam/eicas, radar, cpdlc, gpwrs, etc…) to this “brain” AI computer and let it do its thing. It’s all theoretical as the technology isn’t there yet but there are videos of a robot first officer that flew a 737 on the internet. Full pilotless airplanes are a long ways off but having something like this while the other pilot is resting isn’t out of the realm of possibilities.
That's plausible as an automated IRO, with a human in the bunk.
For anything more robust than that, it probably does not make sense to add the extra complexity of having to identify and manipulate human controls... far better to just network into the system and eliminate multiple potential failure modes.
It would be plausible for full autonomy only if the brain was a full artificial general intelligence (AGI), ie human-equivalent in most respects. Nobody knows how far off that is, and there are very real, and very serious ethical and safety issues associated with going there. To say nothing of certifying something like that, which is inherently non-deterministic... and unlike avionics we're used to you'd also need to certify it's sanity. Ie, how do you know it's not going to spontaneously do something we'd consider irrational, possibly for reasons which make perfect sense to the AGI?