I was in ATC training in the 90s and this was discussed among teachers and ATC personell. The common saying was that pilots would disappear from cockpits before ATC personell were removed, at least from tower control. There are typically three kinds of ATC: Tower control, approach/departure control and area control for controlling planes when cruising. I haven't followed this in years but my impression is that better monitoring equipment allows for fewer area controllers to control bigger areas. I believe area control is the most likely to get automated but this is quite a guess. Approach control is about using radar (or no radar, procedural approach control is a thing) lining up planes to land on a runway. The planes are handed over from approach to tower control when the plane is on final approach. There is also ground control for taxiing on larger airports. But, not least. Do not underestimate the value of having trained personell using radio to great effect. Any belief that modern touch gadgets are better than radio is silly. Humans are also very capable at speaking while performing advanced tasks.
but forget the focus on automating air traffic control, datalink, complex ground IT, remote controls.. That is way to costly and difficult to do in the context of a collection of decentralized legacy systems.
Instead most people are trying to get rid of paper strips (notes used by ATC), and sell complex system that try to automate conflict management.
The hard thing is to improve the UX, the ATC has to communicate with humans (hard even with the highly codified language used), and DO NOT want to solve technical issues, the system has to indicate potential conflicts well in advance but not nag for it at a bad time. They are a lot of human factors to take in consideration and a system well designed with the air traffic controller at the center of it could help a lot.
It's been reported that the elevation of the helicopter was reported as hundreds of feet off. It's unlikely it was just an issue at the specific tower the crash occurred at. If they can't even get accurate elevation data there's no way they'll be able to automate.
I don't know more about ATC, but it looks like a field ripe for disruption and innovation. AI should be able to handle the coordination of flights without the downside of the delays and limitations of the human training pipeline, worker fatigue, and stress - all for less expense. The more I think about it, the more I feel like I could have something tangible at the end of a weekend or two - at least a prototype.
I sincerely hope this is satire (it sure is very HN in nature). "AI" in its current generative incarnation is prone to hallucinations/confabulations that cannot be avoided. In what world is that compatible with a job where a mistake can kill hundreds of people a few minutes or seconds later?
one that you would trust the lives of thousands of humans to every day? It seems unlikely we are anywhere close to a point where we can ensure that any AI won't hallucinate and cause an issue.
Sure, AI can spit out nonsense, and that’s a real concern. But in engineering, we deal with imperfect tradeoffs every day - it’s baked into the job. If we insist on a flawless solution before shipping anything, we’ll never ship. There’s always an optimum where we uphold safety standards without sacrificing forward progress
Could Congress support AI research and innovation by asking AI company CEOs found guilty of overpromising to prove the reliability of their latest technology by flying in AI-controlled airplanes and relying on AI-managed air traffic, instead of using private jets with human pilots and air traffic controllers? /s
I was in ATC training in the 90s and this was discussed among teachers and ATC personell. The common saying was that pilots would disappear from cockpits before ATC personell were removed, at least from tower control. There are typically three kinds of ATC: Tower control, approach/departure control and area control for controlling planes when cruising. I haven't followed this in years but my impression is that better monitoring equipment allows for fewer area controllers to control bigger areas. I believe area control is the most likely to get automated but this is quite a guess. Approach control is about using radar (or no radar, procedural approach control is a thing) lining up planes to land on a runway. The planes are handed over from approach to tower control when the plane is on final approach. There is also ground control for taxiing on larger airports. But, not least. Do not underestimate the value of having trained personell using radio to great effect. Any belief that modern touch gadgets are better than radio is silly. Humans are also very capable at speaking while performing advanced tasks.
Ton of people are working on it,
but forget the focus on automating air traffic control, datalink, complex ground IT, remote controls.. That is way to costly and difficult to do in the context of a collection of decentralized legacy systems.
Instead most people are trying to get rid of paper strips (notes used by ATC), and sell complex system that try to automate conflict management.
The hard thing is to improve the UX, the ATC has to communicate with humans (hard even with the highly codified language used), and DO NOT want to solve technical issues, the system has to indicate potential conflicts well in advance but not nag for it at a bad time. They are a lot of human factors to take in consideration and a system well designed with the air traffic controller at the center of it could help a lot.
It's been reported that the elevation of the helicopter was reported as hundreds of feet off. It's unlikely it was just an issue at the specific tower the crash occurred at. If they can't even get accurate elevation data there's no way they'll be able to automate.
I don't know more about ATC, but it looks like a field ripe for disruption and innovation. AI should be able to handle the coordination of flights without the downside of the delays and limitations of the human training pipeline, worker fatigue, and stress - all for less expense. The more I think about it, the more I feel like I could have something tangible at the end of a weekend or two - at least a prototype.
I sincerely hope this is satire (it sure is very HN in nature). "AI" in its current generative incarnation is prone to hallucinations/confabulations that cannot be avoided. In what world is that compatible with a job where a mistake can kill hundreds of people a few minutes or seconds later?
one that you would trust the lives of thousands of humans to every day? It seems unlikely we are anywhere close to a point where we can ensure that any AI won't hallucinate and cause an issue.
Sure, AI can spit out nonsense, and that’s a real concern. But in engineering, we deal with imperfect tradeoffs every day - it’s baked into the job. If we insist on a flawless solution before shipping anything, we’ll never ship. There’s always an optimum where we uphold safety standards without sacrificing forward progress
2 replies →
I would not want a text generator to "handle" anything responsible for human lives.
Could Congress support AI research and innovation by asking AI company CEOs found guilty of overpromising to prove the reliability of their latest technology by flying in AI-controlled airplanes and relying on AI-managed air traffic, instead of using private jets with human pilots and air traffic controllers? /s