- Question
- Will AI take our jobs?
- Position‹2 of 3›
- Partially: AI will take many jobs, but not all of them
- Argument1 of 2›
AI has already taken many jobs
AI is capable of undertaking lots of human jobs and tasks, but overall humans are still preferred
The argument
The argument accepts the first half of the automation claim and rejects the second. It is already the case, proponents note, that many jobs have been taken: certain factories have successfully replaced 90% of their human workforce, and AI now performs complex tasks that were once treated as beyond mechanisation altogether, such as navigating air transit, performing medical diagnoses and composing new types of music. The point of listing tasks of that kind is that they are not routine. If the frontier of automation were fixed at repetitive manual work it would be easy to bound; that it reaches judgement, diagnosis and composition is what makes the stronger claim seem plausible. And yet, on this view, most jobs are still for humans. What explains that is a pattern in how automation is actually deployed. Even in industries where automation represents a cheaper option than human labour, it takes up only a certain percentage of the available work rather than all of it. Supermarket self-checkouts are the illustration: they are almost always accompanied by human assistants, and by a series of human checkout workers alongside them. The cheaper technology was adopted and the human roles were not eliminated, only redistributed, some of them towards attending to the machines and to the transactions the machines cannot complete. Proponents draw the general conclusion from that example. A cost advantage does not on its own determine how much of a job automation absorbs, because the tasks within a job differ in how tractable they are, and the intractable residue still requires a person. So while automation may look as though it is close to taking over, the pattern to date suggests it will only ever supplement human work rather than displace it wholly. Many jobs will be taken; not all of them will be.
Premises
Counter-arguments
While this may be the case now, what happens when the price of automated work becomes significantly lower than the price of human labour? Why would companies continue to employ humans when self-checkouts are faster, cheaper, and easier?
Rejecting the premises
P1 - This is just the beginning P2 - Human labour will become more expensive while automation gets cheaper P3 - It may not end it straight away, but it could well do in the future