I strongly disagree with this assertion. There's a lot of things we actively don't do this with for all sorts of reasons both moral/ethical, practical/political creepiness/impact. I'll give some...
Where we can make something superiour to humans, we will.
I strongly disagree with this assertion. There's a lot of things we actively don't do this with for all sorts of reasons both moral/ethical, practical/political creepiness/impact.
I'll give some examples:
several European countries don't automate denial of asylum requests and require humans to make independent or oversee all denials of asylum. At the same time, they automate many grants of asylum.
Voter registration could easily be automated. The US doesn't do this federally for various reasons. Many other countries do for other and sometimes the exact same reasons the US doesn't.
Algorithms could create infinitely more fair voting districts based on clearly defined principles/parameters. Politicians don't want those superior and fair districts, so the process isn't automated.
Algorithms are in many cases better at predicting consumer behavior, like whether or not to give someone a loan. Banks largely don't automate housing loans etc. because some of the factors computers choose can be discriminatory or unwanted.
A lot of things (like road tolls) aren't automated in different places because that'd mean people lose their jobs.
Personalized ad systems get creepily accurate. So much so that it's scary, so many companies feed ads you know you're not interested in to avoid the creepiness feel. A famous example here is that Walmart often revealed to couples that they were pregnant before they knew themselves though ads. Walmart intentionally delays/worsens automated ads
Businesses could largely automate customer support though emails/robots on the end of phones. They generally don't go nearly as auto as they could, because consumers hate interacting with robots.
UAV systems require humans to be in the loop to verify launches in many cases where they wouldn't need to be, because the consequences of mistakes are so high you can't trust bots, even though they make fewer mistakes than people in disregarding potential flawed/illegal orders.
Wealth funds could often do much better by letting algorithms/bots manage money. However, most don't do this. There are many reasons why brokers aren't all out of work even though it costs us all a lot of money.
Automation is new. There need to be specific reasons for why automation would be inevitable in all sorts of cases where non-efficiency concerns trump efficiency.
I strongly disagree with this assertion. There's a lot of things we actively don't do this with for all sorts of reasons both moral/ethical, practical/political creepiness/impact.
I'll give some examples:
several European countries don't automate denial of asylum requests and require humans to make independent or oversee all denials of asylum. At the same time, they automate many grants of asylum.
Voter registration could easily be automated. The US doesn't do this federally for various reasons. Many other countries do for other and sometimes the exact same reasons the US doesn't.
Algorithms could create infinitely more fair voting districts based on clearly defined principles/parameters. Politicians don't want those superior and fair districts, so the process isn't automated.
Algorithms are in many cases better at predicting consumer behavior, like whether or not to give someone a loan. Banks largely don't automate housing loans etc. because some of the factors computers choose can be discriminatory or unwanted.
A lot of things (like road tolls) aren't automated in different places because that'd mean people lose their jobs.
Personalized ad systems get creepily accurate. So much so that it's scary, so many companies feed ads you know you're not interested in to avoid the creepiness feel. A famous example here is that Walmart often revealed to couples that they were pregnant before they knew themselves though ads. Walmart intentionally delays/worsens automated ads
Businesses could largely automate customer support though emails/robots on the end of phones. They generally don't go nearly as auto as they could, because consumers hate interacting with robots.
UAV systems require humans to be in the loop to verify launches in many cases where they wouldn't need to be, because the consequences of mistakes are so high you can't trust bots, even though they make fewer mistakes than people in disregarding potential flawed/illegal orders.
Wealth funds could often do much better by letting algorithms/bots manage money. However, most don't do this. There are many reasons why brokers aren't all out of work even though it costs us all a lot of money.
Automation is new. There need to be specific reasons for why automation would be inevitable in all sorts of cases where non-efficiency concerns trump efficiency.