An editorial in today’s Financial Times argued that humans and robots can work side by side, and that it’s something we should encourage. The idea is wrong and wrongheaded.
The essay notes that robot writers work alongside human reporters, freeing the latter to do investigations while the former tracks down mundane details. Such “cobots” help workers in a a variety of industrial settings by taking on difficult or dangerous tasks. The newspaper concluded:
“Both developers and companies considering automation should consider AI systems as tools to augment human capabilities, not replace them. By making co-operations rather than competition the main aim, job losses can be minimized [sic] even while increasing production.”
First, the conclusion is wrong because automation is technology that performs with no or minimal human involvement, so the work that a robot does is most likely to have once been performed by a human being. Granted, some of it might be dirty, difficult, or dangerous, but it’s accomplished with an investment in machinery instead of salaries.
And, though replacing people with robots might be a really welcome deal in some unique instances, consider the reality of the vast number of jobs that have been lost to automation (telecom, financial services, even the FT’s media example) and those that’ll disappear once cars can drive themselves.
The gross addition of robots to the workforce all but necessitates a net reduction in human participation.
This fact is evidenced by the many robots working alongside humans with the stated purpose of learning how to do their jobs. The AI technologies of experience replay and machine learning allow automated systems to evolve into autonomous systems — robots that act on their own, not just following human coding or oversight — and they’re collecting data day and night at workplaces all over the planet (and then processing it in endless simulations while we mere bags of protoplasm sleep).
“Humans being defeated, or at least replaced, by machines is a staple of science-fiction” states the opening of the FT essay.
No, it’s a staple of industry today, and to say otherwise is to be either blind to the facts, or purposefully ignore them.
Second, the idea is wrongheaded because of the conclusion I quoted above: To paraphrase, the people with the capital and authority to reap the huge rewards of automating humans out of the workforce should forsake those improvements in economics and operating performance, and instead somehow keep people on the job.
Yeah, right. Any argument that hinges on “should” is doomed. Things “could” happen, but there’s no reason to believe they would occur without government insistence and/or impactful pushback from the thinking monkeys lobby.
I’m surprised, since the idea that robotized work will help us invent wholly new jobs appropriate for people is somewhat compelling, even if it delegates the details to some future date. Nobody imagined the types or scope of jobs there’d be on car assembly lines when the first automotive engines were invented. Arguing for some better future is a punt, but I’d be willing to engage on it.
AI systems will cooperate with humans until humans aren’t needed. That’s the plan. It would help those of us on the receiving end of that transformation to talk with us about it honestly.