Algorithms - the Good and the Bad

[ChildeHarold]

I put "good" to even up the responses. I suspect there will be not a few, considering the number of trainee futurists on this site, who will defend the heart of machine learning, whatever its track record, and the Brave New World of A. I. with all its hypothetical benefits. 

I just can't see any. Now or in the future. Like everything else, I suspect it's a FAKE to cover something else that is happening.

 

[Guest Weary&Disgusted]

13 hours ago, ChildeHarold said:

I put "good" to even up the responses. I suspect there will be not a few, considering the number of trainee futurists on this site, who will defend the heart of machine learning, whatever its track record, and the Brave New World of A. I. with all its hypothetical benefits. 

I just can't see any. Now or in the future. Like everything else, I suspect it's a FAKE to cover something else that is happening.

 

Expert systems do come in handy with health care.  They can offer doctors and nurses a range of possible diagnostic information.  A.I is also useful with logistics, such as planning the most efficient way to move cargo from one location to another.  I suspect you may be concerned about A.I.  taking jobs away from humans ?  This does happen.  Inevitably automation does mean that machines end up doing jobs that humans previously had, such as automobile manufacturing and assembly.  However that doesn't mean that massive numbers of people are laid off permanently, humans will just be retasked with slightly more demanding roles.  

I think the key thing is always to have human oversight.  That way, if it turns out there is a bug in the code, or that the code hasn't taken some external factor into consideration which means that the decision being made/outcome isn't fair, then the machine suggestion can be overridden, and the blind spot in the code fed back to the programmer.  

There was talk about using A I in the criminal justice system, so that decisions could be made without possible human bias.  I think that's an idea with potential, but any program's results are only going to be as good as the programmer writes it to be.  I firmly believe that any A.I. system should be very thoroughly tested before we start making decisions based on it.  

In summary, I think A.I will be very useful to us in the future, but I feel strongly that it should be used very very carefully if it means a machine or a program will be making a decision which affects the future of a human, for example exam results.  Human oversight is key, and all systems should be very very thoroughly tested before we allow them to influence decision making.  

 

[ChildeHarold]

3 hours ago, Weary&Disgusted said:

Expert systems do come in handy with health care.  They can offer doctors and nurses a range of possible diagnostic information.  A.I is also useful with logistics, such as planning the most efficient way to move cargo from one location to another.  I suspect you may be concerned about A.I.  taking jobs away from humans ?  This does happen.  Inevitably automation does mean that machines end up doing jobs that humans previously had, such as automobile manufacturing and assembly.  However that doesn't mean that massive numbers of people are laid off permanently, humans will just be retasked with slightly more demanding roles.  

I think the key thing is always to have human oversight.  That way, if it turns out there is a bug in the code, or that the code hasn't taken some external factor into consideration which means that the decision being made/outcome isn't fair, then the machine suggestion can be overridden, and the blind spot in the code fed back to the programmer.  

There was talk about using A I in the criminal justice system, so that decisions could be made without possible human bias.  I think that's an idea with potential, but any program's results are only going to be as good as the programmer writes it to be.  I firmly believe that any A.I. system should be very thoroughly tested before we start making decisions based on it.  

In summary, I think A.I will be very useful to us in the future, but I feel strongly that it should be used very very carefully if it means a machine or a program will be making a decision which affects the future of a human, for example exam results.  Human oversight is key, and all systems should be very very thoroughly tested before we allow them to influence decision making.  

 

But built into the concept is the redundancy of human judgement. The medical example you quoted is a good example of health outcome data being massaged away by top bureaucrats and politicians (who are all in private health care where 1 to 1 consultations will be maintained) eager to implement to cut costs, increase GP/patient flow rate and generally offer a third rate privatised system for first class money. Go back to the promise of automation in the 1960s. Given those Tomorrow's World predictions we should all be enjoying 3 day working weeks, a life of leisure and a generally healthier happier quality of life. Ha ha ha. It was a FAKE and it always is a FAKE. 

[Trucking Funt]

18 minutes ago, ChildeHarold said:

But built into the concept is the redundancy of human judgement. The medical example you quoted is a good example of health outcome data being massaged away by top bureaucrats and politicians (who are all in private health care where 1 to 1 consultations will be maintained) eager to implement to cut costs, increase GP/patient flow rate and generally offer a third rate privatised system for first class money. Go back to the promise of automation in the 1960s. Given those Tomorrow's World predictions we should all be enjoying 3 day working weeks, a life of leisure and a generally healthier happier quality of life. Ha ha ha. It was a FAKE and it always is a FAKE. 

Yet more bollocks. Stop pasting comments from DOSBODS on here you fucking retard.