The CBC’s Don Pittis is back with another article on automation: These Canadians are helping the world become replicant ready.
Generation R, a startup company that helps organizations prepare for the robot invasion, recently completed a study for Technical Safety BC, a self-funding organization charged with licensing and inspecting the safety of technical installations in the province. The technical authority recently began incorporating machine learning — the basis for modern artificial intelligence — into its system for deciding where to get the best bang for its buck in the use of its limited inspection staff.
Seems like a pretty mundane application of AI.
The job of Generation R was to spot where the new AI system was likely to encounter problems with the human-centred task. Workers at Technical Safety BC were worried that the new automated prediction algorithm would create ethical problems, missing or misjudging risks or stealing jobs rather than helping workers to do their jobs better.
The first concern, the accuracy of the AI, is important to test. The second, that the AI would “steal jobs” seems silly, considering that the AI is being used to allocate scarce human resources to inspections. The AI is being used to improve the efficiency of the human inspectors by targeting the most important problems. This might lead to the need for fewer human inspectors, but the AI is not stealing their jobs.
In a Media Market store In Switzerland last week, customers appeared disconcerted by a robot named Paul that can help find products on the shelves.
Given how hard it is to find help in large stores today, it’s hard to see why anyone wouldn’t be happy with such a service. Automated price checkers and bookstore kiosks that give you stock and location are limited forms of the same kind of automation, and they are vast improvements over searching for human help in an understaffed store.
Experts say automation is best when it is stealing jobs in what they call the three Ds, those that are dull, dirty or dangerous.
Dull jobs like factory assembly are often repetitive, and are well suited to automation. Dangerous jobs like inspecting nuclear reactors or bomb disposal are high paying, and therefore there is a bigger benefit to automating them. Today’s machines aren’t as good at dealing with dirty, chaotic environments, but, like dangerous jobs, dirty jobs have to pay a premium, so automating them when possible has a good payoff.
According to Elizabeth Croft, the idea that a higher minimum wage will take away dull and repetitive … low-wage jobs is inevitable. The trick is to make people happy about it. To maintain our standard of living we actually have to embrace this kind of technology.
Higher wages do increase the pressure to automate, particularly when we are competing with countries that have far lower labor costs, and against others who are aggressively automating to reduce them. The question is, whose standard of living will be maintained by automation? Certainly not those who lose their jobs due to it. A higher minimum wage doesn’t help if you’re out of a job. See my post Raising the Minimum Wage Can Backfire.
The way to do that is to be sure there are still plenty of good quality, complex and interesting jobs for human labour that, so far at least, only humans can do. That seems to be working out. Where robots are good is reliability, repeatability, the heavy lifting, able to untiringly do dumb tasks — pick-and-place pick-and-place, they can do that over and over again.
The key words here being “so far”.
You want to focus your labour to those high-value activities where there needs to be logic under uncertainty. Moving people into those higher-value, higher-wage jobs is the only way to increase Canadian productivity, leaving the bad jobs for the computers and robots.
But not everyone will be capable of doing such jobs. People are not interchangeable. Human intelligence exists on a bell curve. See my post Automation is not the Industrial Revolution and the site it references, Modern IQ ranges for various occupations.
“To be able to do that effectively there is a point where people and robots have to come together to really obtain that full value of that transition,” says Croft.
For the foreseeable future, robots will be mere tools, albeit very sophisticated ones. They will need to be usable, configurable, and safe.
Essentially, the robots have to be constructed and used in a way that makes people happy and comfortable.
This doesn’t follow. Not causing turnover in human employees is important, but overall cost savings will be paramount.
There have always been a few who were incapable of holding a productive roll in society. Now, for the first time, automation raises the specter of a world where more than a small minority are unemployable. If we aren’t far more careful with our spending, the welfare state may collapse under the weight of a whole class of people who have no place in the modern economy. As I’ve said before, things will thankfully change gradually, but the changes have the potential to reshape our civilization in ways it has never been before.