It’s worrying enough that faulty instruments and sensors can cause your pilot to just manually crash, let alone the autopilot.
If you fly commercially then you’re flying in an airplane that flies mostly autonomously. The vast majority of the pilots interaction is during takeoff and landing. The rest of the time it’s pretty much flying itself.
One huge difference between commercial aircraft & cars is that the air traffic is highly regulated. There are specific corridors & altitudes, distances between aircraft that must be maintained, and ground based controllers are monitoring and directing everything. It’s not even remotely like the free for all on the highways.
ChaoticEntropy@feddit.uk 5 months ago
c0mbatbag3l@lemmy.world 5 months ago
There’s still a pilot onboard if they need to takeover, what this article is describing is like having a ground based drone operator.
“To address the pilot shortage”, ok so how long until each ground based pilot is flying multiple aircraft?
WHYAREWEALLCAPS@kbin.social 5 months ago
Since a pilot is mainly only needed for take off and landing 99.9% of the time, I'd bet not much longer. Getting from an to the ground are the hard parts. Even then, during optimal conditions, you probably don't even need the pilot, so it's likely you'll get pilots monitoring multiple aircraft from take off to landing and only interceding when an emergency arises. In time it'd probably become something they can do from their own homes. Then after they've collected enough data and trained the AI with it, they'll be able to do away with the human pilot altogether.
Any task humans do that don't rely on smell or taste you can train an AI to do. The problem is no longer if you can make a machine smart enough to do it, the problem now is do they have enough data to train it. By making pilots remote they'll be able to gather that data easily. Once we develop sensors that can perfectly mimic taste and smell, humans will only be needed for manual labor.
c0mbatbag3l@lemmy.world 5 months ago
You have a grossly oversimplified understanding of human input during a flight. Once at altitude and cruising autopilot can be used, but a certain amount of hours actually flown as well as takeoff and landing are required, and AI isn’t near good enough to take-over during an emergency the way a human operator can react to the situation.
Also manual labor won’t be a thing for long, Boston Dynamics is making progress hand over fist when it comes to quadruped and biped robotics. The moment we train models to handle labor jobs they’ll be out the door as well.