2 minds are better than one. It's a pretty simple rule of thumb that has been time tested, but it's the hype behind driverless vehicles that I think is driving much of this.
The problem is some politicians and others who probably haven't spent much time in a flight deck recently are buying the hype. I do worry that they will make financially driven decisions that will compromise safety.
Even driverless cars have significant hurdles, and I think some politicians, engineers, and regulators have rushed its technology.
https://www.bloomberg.com/news/artic...ash-in-Arizona
Unfortunately it takes an accident for people to start appreciating the limitations of technology.
The problem is with aircraft accidents consequences are far worse when things go south... even on a cargo aircraft. Which is why I think its good ALPA and others work to slow this research way, way down, before someone gets the crazy idea to try this thing on 121 ops and someone gets hurt.
I think huge problems remain even from a single pilot perspective
Security
- One person at the controls of a potential WoMD. (German Wings)
- Hackable interfaces (all remote tech will have vulnerabilities, if they don't today they will tomorrow)
Redundancy
- We take flying aircraft for granted because we do it all of the time, but just talk to a new student and one will discover it is not a "simple learning experience" a lot of intuition "experience" goes into dealing with (People, Equipment, Weather, Equipment Failure, regulation, and other components)
- Single pilot ops are much more prone to fatigue (judgement errors)
Judgement
-Common sense isn't so common, and computers rank near the bottom of the rating scale when it comes to having it, which includes AI.
Technology
- Equipment failure...
- The AI of today are essentially graphics processors that use inputs (experience) to slowly derive the weightings of a formula to learn the right outputs for a given input.
The problem is as most humans know, no formula can accurately model the proper output in all cases, no matter how careful weighted the components of the formula are, because no formula or sets of formula will be comprehensive enough to compare with humans. Our brains were developed over long, long periods of time, in ways engineers can't even dream of right now with current tech. (Just the placement of our neurons, or the folds in our brains are unique) Graphics processers are well.... very rigid.
Our judgement in the flight deck stems from a lot more than just inputs and outputs from flying. It comes from life experience, of simple stuff we take for granted.
If it turns out AI has been developed good enough to exercise the necessary judgment in our crazy world, to fly aircraft, then we might as well have robots for senators or CEOs.
In the end people need to work with people, which is why 2 is such a strong team. They both bring different perspectives, and often can act synergistically when it comes to good judgment, while providing a level of redundancy that R2D2 simply can't match.