Considering the current advances of AI how likely do you think an AI might replace you in your job?

[Christine Wheelwright]

Modern airliners can almost fly themselves. After the flight plan has been inputted into the FMS the plane can practically fly itself. And if I'm not mistaken, in certain situations autonomous landings do occur. Computers have become an integral component of modern aircraft. Modern fighter jets wouldn't be able to achieve the performance they have without modern computer systems (controlled by wire). With that being said, when diversions or emergency scenarios occur you want a well trained (and rested) human being at the controls. Technological innovation is progressing faster than ever, the time between one innovation and the next becoming smaller, I think it's only a matter of time before AI would be able to act as an actual pilot. However, I doubt it'll be anytime soon.

Yes, exactly. There is a temptation to think that the technology is almost there, just because it copes 99.9% of the time. But that other 0.1% is crucial, and can be a thousand times harder to achieve than the first 99.9% was.

 

[CrazyKB]

Yes, exactly. There is a temptation to think that the technology is almost there, just because it copes 99.9% of the time. But that other 0.1% is crucial, and can be a thousand times harder to achieve than the first 99.9% was.

Have a Tim Hortons on me.

 

[Karn's Return]

The twist:

I'm unemployed and just surviving.



I voted 100% certain.

 

[Wayne Mack]

Modern airliners can almost fly themselves.

Let me unpack that statement. What autopilot systems do, and do very well, is control the airplane under the common situations that they have been explicitly programmed to handle. If they encounter a different situation, then they should turn over control to the pilot, but even here the determination that there is an unusual situation needs to be explicitly programmed in. A dramatic example from a few years back is the 737-Max autopilot (not AI, but still a very complex virtualization system) fought the human pilots and took action that would deliberately crash the plane.

Similarly, machine learning systems do very well when handling the exact scenarios that they have been trained on. Where issues arise is in handling situations and scenarios that they haven't been trained on or haven't been sufficiently trained on. These systems typically produce a confidence rating, a weighted scoring of multiple variables and translate that into a Yes, No, or Could Not Determine result. It is this middle area that causes unknown problems to occur. A Yes may be a false positive while a No may be a false negative, but as the systems are designed to minimize human support, they often aggressively push decisions into the Yes or No area. An example of this has be facial recognition, where the training data matches has been dominated by white, male images. The systems do quite well in those cases, but the matching algorithms reliability quickly deteriorates for other populations. It then becomes a chicken and egg problem to increase training data with representation of images that were unknowingly omitted from training data.

I suggest that AI is best suited for augmenting human perception and increasing human engagement in tasks. Decreasing human engagement makes it less likely for a person to be able to override bad decisions by AI and to be able handle situations that the AI does determine are beyond its capabilities.

 

[CrazyKB]

Let me unpack that statement. What autopilot systems do, and do very well, is control the airplane under the common situations that they have been explicitly programmed to handle. If they encounter a different situation, then they should turn over control to the pilot, but even here the determination that there is an unusual situation needs to be explicitly programmed in. A dramatic example from a few years back is the 737-Max autopilot (not AI, but still a very complex virtualization system) fought the human pilots and took action that would deliberately crash the plane.

Similarly, machine learning systems do very well when handling the exact scenarios that they have been trained on. Where issues arise is in handling situations and scenarios that they haven't been trained on or haven't been sufficiently trained on. These systems typically produce a confidence rating, a weighted scoring of multiple variables and translate that into a Yes, No, or Could Not Determine result. It is this middle area that causes unknown problems to occur. A Yes may be a false positive while a No may be a false negative, but as the systems are designed to minimize human support, they often aggressively push decisions into the Yes or No area. An example of this has be facial recognition, where the training data matches has been dominated by white, male images. The systems do quite well in those cases, but the matching algorithms reliability quickly deteriorates for other populations. It then becomes a chicken and egg problem to increase training data with representation of images that were unknowingly omitted from training data.

I suggest that AI is best suited for augmenting human perception and increasing human engagement in tasks. Decreasing human engagement makes it less likely for a person to be able to override bad decisions by AI and to be able handle situations that the AI does determine are beyond its capabilities.

The crash of Boeing 737 Max appeared to be caused by the MCAS putting the plane into a nose dive due to erroneous data from the AoA sensors. From what I remember, the MCAS 'thought' that the aircraft was about to enter a stall, it compensated by dropping the nose (resulting in the dive). However, from what I understand, the autopilot wasn't engaged at the time, since one of the criteria required for the MCAS to activate is that the plane must be manually flown. That same scenario has happened to pilots, where they receive conflicting data from different systems (because one was wrong), since they had no visual references to go by (flying a night for example), they had to guess which data was accurate, unfortunately some choose wrong resulting in a crash.

Our brains are also prone to false positives (type one errors). Our brains (fusiform gyrus) also create meaningful interpretations based on erroneous patterns (patterns that don't exist). For example, seeing a 'face in the clouds' (Pareidolia).

One difference between computers and brains is memory degradation. If we don't process data frequently we tend to forget it. We can 'jog our memories' by reading about something we've forgot for example, however it's difficult to recall things we haven't thought about in a long time without some stimulus to help us remember, most of the time our memories aren't entirely accurate either. A computer doesn't have that problem. The one key difference between humans minds and artificial ones is the latter's inability to 'think outside the box'. When AI can take what has been programmed into it (akin to human learning) and come up with solutions based what it 'knows', I think AI's processing power will go far beyond what our brains are capable of. AI has the potential, at least in theory, to sort though enormous amounts of data extremely quickly and form novel solutions to problems. I also believe that we'll fuse our biology with technology in the far future, the so called 'singularity' (not to be confused with how that term is used in physics).

 

[CultureCitizen]

Modern airliners can almost fly themselves. After the flight plan has been inputted into the FMS the plane can practically fly itself. And if I'm not mistaken, in certain situations autonomous landings do occur. Computers have become an integral component of modern aircraft. Modern fighter jets wouldn't be able to achieve the performance they have without modern computer systems (controlled by wire). With that being said, when diversions or emergency scenarios occur you want a well trained (and rested) human being at the controls. Technological innovation is progressing faster than ever, the time between one innovation and the next becoming smaller, I think it's only a matter of time before AI would be able to act as an actual pilot. However, I doubt it'll be anytime soon.

Office tasks will get automated first because they don't really require a physical robot, they only require an AI to move data around. I think most of white-collar jobs will be automated in the next 12 years. On the other hand, jobs that require interaction with the environment require a robot to replace a human, this requires more training ( due to all the noise and variation coming from the environment) and an additional cost ( the actual robot). It will probably take 30 to 50 years before these kinds of jobs are replaced by machines, mostly because the production of robots has to be ramped up.

 

[CrazyKB]

Office tasks will get automated first because they don't really require a physical robot, they only require an AI to move data around. I think most of white-collar jobs will be automated in the next 12 years. On the other hand, jobs that require interaction with the environment require a robot to replace a human, this requires more training ( due to all the noise and variation coming from the environment) and an additional cost ( the actual robot). It will probably take 30 to 50 years before these kinds of jobs are replaced by machines, mostly because the production of robots has to be ramped up.

What do you mean by 'interacting with the environment'? Also, wouldn't robots be cheaper in the long run, since they don't require breaks, their output being far beyond what we're capable of and they don't have to be paid. The only costs would be their construction and maintenance. If AI replaces most jobs what do you think would happen to the economy?

 

[Vertigo]

What do you mean by 'interacting with the environment'? Also, wouldn't robots be cheaper in the long run, since they don't require breaks, their output being far beyond what we're capable of and they don't have to be paid. The only costs would be their construction and maintenance. If AI replaces most jobs what do you think would happen to the economy?

Eventually yes and it will of course happen. Theoretically we eventually hit the 'post-scarcity' economy but realistically that will never happen whilst most wealth is owned by so few. But I will stop there before we drift into forbidden territory!

 

[Cthulhu.Science]

To broaden to Automation - not simply the specific of whatever AI is - Automation has been coming for everyone's job for a very, very long time.

Consider airline operations crew. There used to be 3 people in the cockpit. Now there are two, but airlines in the US are fighting to reduce that to one. There are fewer checkout clerks at large stores, because scanning is so much faster than typing each number. And now there is "self-checkout" where a single clerk monitors upward of six registers. Or the very famous Amazon warehouse robots. The job "typist" is gone. How many carpenters lost their jobs to power tools and air-nailers? Or going back a bit, when did wind and water mills take over all the work of grinding grain into flour? All of those man-hours of work simply gone to automation.

In fact we could add to this thread, or start an all new thread simply listing jobs that were made obsolete or greatly reduced by automation.

In a hundred years, will people scoff that actual people used to do your job? I hope they do for mine.

 

[Cthulhu.Science]

Office tasks will get automated first because they don't really require a physical robot, they only require an AI to move data around. I think most of white-collar jobs will be automated in the next 12 years. On the other hand, jobs that require interaction with the environment require a robot to replace a human, this requires more training ( due to all the noise and variation coming from the environment) and an additional cost ( the actual robot). It will probably take 30 to 50 years before these kinds of jobs are replaced by machines, mostly because the production of robots has to be ramped up.

Most white collar jobs have been lost to automation (computers mainly) during the last 50 years. Typing pools being the most obvious. There were a lot of people whose job it was to manage paper files. Jobs that no longer exist.

 

[Cthulhu.Science]

What do you mean by 'interacting with the environment'? Also, wouldn't robots be cheaper in the long run, since they don't require breaks, their output being far beyond what we're capable of and they don't have to be paid. The only costs would be their construction and maintenance. If AI replaces most jobs what do you think would happen to the economy?

Lots of people likes to personify automation in the forms of Robots or Artificial Intelligence. In a variety of ways, machines and automation have been eliminating jobs for a very long time. Consider this article.

Oxen, the ‘robots of the late Neolithic’ jump-started economic inequality​

 

[CultureCitizen]

What do you mean by 'interacting with the environment'? Also, wouldn't robots be cheaper in the long run, since they don't require breaks, their output being far beyond what we're capable of and they don't have to be paid. The only costs would be their construction and maintenance. If AI replaces most jobs what do you think would happen to the economy?

A translator simply gets text and produces a translation.
It has no interaction with the physical environment.
A car's automatic pilot requires input from the physical environment : the street , pedestrians an cars.

The only costs would be their construction and maintenance.

Mostly the energy consumed by the robots.

If AI replaces most jobs what do you think would happen to the economy?

I'll answer that in a private message to avoid having the thread deleted.

 

[CultureCitizen]

Lots of people likes to personify automation in the forms of Robots or Artificial Intelligence. In a variety of ways, machines and automation have been eliminating jobs for a very long time. Consider this article.

Oxen, the ‘robots of the late Neolithic’ jump-started economic inequality​

I agree, although countries with aging populations ( Germany, Japan, France) have the opposite problem.
Generally speaking there is a perpetual scarcity of jobs.

 

[CultureCitizen]

Eventually yes and it will of course happen. Theoretically we eventually hit the 'post-scarcity' economy but realistically that will never happen whilst most wealth is owned by so few. But I will stop there before we drift into forbidden territory!

We can go either way: Cyberpunkish or Space Opera Utopia.

 

[Cthulhu.Science]

Cleaning robots at BKK Airport, Thailand. Taking jobs in a place where labor is very, very cheap.

 

[Cthulhu.Science]

We can go either way: Cyberpunkish or Space Opera Utopia.

Kurt Vonnegut "Player Piano" is another option, and seems more like the direction we are heading.

 

[Cthulhu.Science]

We could delve into a lot of Vonnegut for visions of the future. Somehow they never seem to end well.

 

[AllanR]

ars technica just published an article on this topic. Generative AI set to affect 300 million jobs across major economies

 

[CrazyKB]

Eventually yes and it will of course happen. Theoretically we eventually hit the 'post-scarcity' economy but realistically that will never happen whilst most wealth is owned by so few. But I will stop there before we drift into forbidden territory!

Without venturing into 'forbidden territory' ... If AI replaced most jobs people wouldn't make money and the economy would collapse.

We can go either way: Cyberpunkish or Space Opera Utopia.

Perhaps we'll figure out how to convert energy into matter (since they're the same thing) like they do in Star Trek. I believe that any spacefaring civilization capable of traveling to other solar systems would have to avoid, or at least get past, a dystopian future, if they don't they'd go extinct long before they could achieve (advanced forms) of space travel (which isn't a novel idea, it's found in many musings, including the Drake Equation).

 

[Cthulhu.Science]

Without venturing into 'forbidden territory' ... If AI replaced most jobs people wouldn't make money and the economy would collapse.



Perhaps we'll figure out how to convert energy into matter (since they're the same thing) like they do in Star Trek. I believe that any spacefaring civilization capable of traveling to other solar systems would have to avoid, or at least get past, a dystopian future, if they don't they'd go extinct long before they could achieve (advanced forms) of space travel (which isn't a novel idea, it's found in many musings, including the Drake Equation).

An example of a post-scarcity economy already exists. Qatar. Where the In-Group has everything and the Out-Group has nothing, not even citizenship.