Like crewless, zero crew. There isn’t even any flight attendants.
Pilots are just an AI Autopilot and flight attendants are all just robots.
Like crewless, zero crew. There isn’t even any flight attendants.
Pilots are just an AI Autopilot and flight attendants are all just robots.
As a pilot, not anytime soon, and not just because I’m worried about my job. Like cars, you can automate 95% of flying pretty easily and for the most part, we already do. But also like cars, that last 5% is several orders of magnitude more difficult.
But cars have a big advantage over planes in automation, if the computer gets totally confused, it can pull over, stop and let the driver figure things out. A plane can’t stop flying without hitting the ground so the computer can’t give up in an edge case. There’s also a different standard for safety. A few dozen teslas slam into walls and not many people care outside of immediate families. 70 people die in a plane crash and it’s international news for months.
I figure it’ll happen, but not anytime soon. And zero pilots is way more feasible than one pilot. And no way in hell can a robo flight attendant manage a cabin in normal operations let alone an emergency, I don’t think that part will ever happen unless we go full synth.
Totally unrelated to the thread, but this reminded me of Special To Hollywood, a really interesting short radio play from the 40s, about a plane that mysteriously becomes suspended in the air while the passengers and crew freak out trying to comprehend what’s happening. It’s a sort of surrealist horror story that I think was actually pretty ahead of its time.
Somehow, as a kid I ended up with a collection of cassette tapes of these Lights Out radio shows. They were all pretty basic, Twilight Zone-y stories, but this one always stood out to me. You made me remember that, so I just wanted to share this weird little story.
Exactly, and accident in aviation are never because of one factor but rather multiple of them, and Murphy’s law.
Handling an engine flame out is ok. But in reality it might be an engine flame out + icing conditions + OVC008 + non precision approach + windshear on short final.
AI pilot would have quit the seat a long time ago, leaving the self loading flesh bags at the rear alone.
ATC here and the only time we consider an aircraft safe is when it’s at the airport. That’s why slots exists. As annoying as they can be, better have the aircraft on ground at DEP rather than holding to wait for the thunderstorm to pass.
Wanna dogpile onto this comment to add that we can’t even automate robots to mow lawns by themselves. The ones you can buy that do any part of the job well at all require GPS, and also require manual intervention or remote piloting for even getting the bot back to its charging station. I work for a corporation that automates machinery like this and sells it to the US government, which advertises its products as automated or crewless, but actually requires somebody at the helm of a software suite to manually adjust and operate the bots at any given moment during their operations. How the hell are people expecting cars and planes to automatically get you to your destination? Imagine your “crewless” vehicle being piloted by some dude in an office somewhere in your country instead of someone actually being at the wheel. Does that make any kind of sense? Would you trust the delay in instructions? What happens when your vehicle can’t receive any outside connections?
Some level of complex “Autonomous” everything, from now until the foreseeable future, will always have a human in the pilot seat. 100% automation is impossible for us right now with our current level of technology.
The reason is more than just that the last few % points of automating is the most difficult hurdle, though I really agree with that part. It’s that automation can’t account for improvising, adapting, innovating. Automation can’t do on-demand problem solving. Space probes on the Moon and Mars can’t unflip themselves when they get stuck. Programmed machines can only do what they’re programmed to do. We’re beyond anything somewhat complex getting 100% automated any time soon.
Accounting? Helpdesk support? Labor that is repetitive and doesn’t require much ingenuity will get automated fast.
Heavy machinery? Art? Transportation? Medical care? We’re 100+ years a way from completely unmanned complex tasks. People eat up the sci-fi marketing garbage without really interacting with or testing the claims being made.
Why is zero more feasible than one?
A single pilot would need the ability to control the plane and override automation and that’s a very dangerous thing.
A single task saturated human in a stressful situation is more of a liability than a benefit. A single pilot can fixate on problems and try to solve them, even if they’ve totally misidentified the situation. Add another set of eyes and you naturally slow things down and handle situations better. The Air France crash in ’09 is a decent example of how one person can totally misinterpret a problem and then the remedy caused a crash, the other crew members were able to figure out what was happening, but it was too late to recover. More crew are an additional chance of success.
Then there’s the whole one pilot is a single point of failure problem. An incapacitated pilot is a fairly straightforward problem, but what about a '15 Germanwings type of situation where the pilot tries to intentionally crash?
Ty for the write up! Makes sense when you put it like that. I’d still trust one human, warts and all, over some ai cooked up by some tech bro. But two humans definitely seems best.