Uncommon Descent Serving The Intelligent Design Community

Finance prof: Artificial intelligence does not threaten complex jobs

arroba Email

Image result for robot public domain image Not any time soon, according to an analyst at Bloomberg:

It’s important to note that machine learning hasn’t yet made its mark on the economy — to paraphrase economist Robert Solow, you can see the machine learning age everywhere but in the economic statistics. Employment levels have returned to healthy levels, and there’s no evidence that machines are taking many of our jobs yet.

The authors generally don’t envision a world of full automation, with machines replacing humans at every step of the production process. Instead, they see machine learning being deployed selectively at some nodes of the value chain where data is plentiful, leaving human judgment to focus on the rest. Though “judgment” is a fuzzy word, Agrawal et al. basically identify two cognitive tasks in which humans will beat intelligent algorithms for the foreseeable future — making predictions based on small data samples, and identifying what constitutes success and failure. Humans are still better at knowing what they want, and at modeling the underlying structure of how the world works.
Noah Smith, “Artificial Intelligence Still Isn’t All That Smart” at Bloomberg

News at Linked In:

Note: My local WalMart is engaged in a so-far disastrous effort to get people to use automated checkouts instead of lining up in front of cashiers we have known and chatted with for years. Many of the cashiers are flapping around the machines to steer people there who don’t want to be steered (because no one asked for this innovation anyway), while the lineups grow longer in front of the cash desks.

A lot depends on whether an innovation is perceived as solving a problem — Is the situation it addresses is even perceived as a problem? – O’Leary for News


See also: Robert J. Marks: Screenwriters’ jobs are not threatened by artificial intelligence Unless, of course, the public starts preferring mishmash to creativity. An AI-generated film is not an altogether new idea. Rule-based expert systems were used to write short plays over a half century ago, in the early 1960’s. Then, as now, don’t expect creativity. That is not what AI does.

Note: Dr. Marks is one of the authors of Introduction to Evolutionary Informatics.

Do children trust robots too much? Maybe, but more study is needed, say researchers. Children could easily give in to peer pressure from other children to give an incorrect answer in place of a correct one. How much difference it makes that the pressure is supplied by a robot would surely depend on how the child is taught to see robots.

Why do we think money is real? It is actually a constantly shifting network of agreements to trust others. Maria Bustillos, editor of Ethereum’s culturemag, Popula, asks us to think about just what money is before we make up our minds about Bitcoin.

In kind of reverse order, I use the automated checkout at Walmart all the time, although there is always a Checkout Helper there to point out where the slot to "insert cash" and such are. But it is EXACTLY Complex Jobs that AI threatens. Piloting an aeroplane is one of the most complex jobs around. Autopilot systems, which include automatic landing, have been around for decades. And the ONLY reasons there are still humans in the cockpits is because: 1) military jock pilots LIKE flying the planes, 2) commercial airlines like having humans in the cockpit so the crash can be blamed on Pilot Error instead of Software Malfunction. (Lawyers can SUE for product failure; they can't sue for stupid humans.) Any number of things that used to require a roomful of degreed engineers with slide rules are now done by a sub-routine in some general purpose engineering program. ALL of the fancy bits of Math (trig functions, calculus, ares or volumes of odd shapes) are done by software. So is statistical analysis. Other than Teaching, I'm not sure what a degree in Mathematics would be used for. AI still has trouble with Art. AI can PLAY music just fine. AI attempts to COMPOSE music have generally failed, probably because the carbon-based fleshy units can't explain how Creativity works. But a friend explained to me some decades back that the reason Opera and Ballet and Orchestras are still around is NOT because they're especially entertaining. They're still around because they're EXPENSIVE, RIDICULOUSLY expensive. And Rich People, who can't think of enough things to spend their money on, can be convinced to PAY ridiculously expensive prices for tickets. If Rich People suddenly decide that "art" is a waste of their money, opera will be right next to Gregorian Chant in the back of the warehouse. And the same is true for any number of things that are "hand-crafted" instead of produced by intelligent machines with REALLY precise tolerances. Someone showed me a picture once of The Future of Human Work. It showed a man sweeping the floor of a large room in which FULLY automated machines were weaving cloth. And the reason to hire the man was because at that point it was still slightly cheaper to hire him instead of using a REALLY smart and efficient automated sweeping machine. Or a better example. Up into the post-WW2 period in the US, there was little automation in farm labor (hoeing weeds between plants, hand-picking cotton, etc.). But the reason this was so is because the farmers had consistently convinced the politicians to EXCLUDE "farm labor" from Minimum Wage Laws. As soon as the laws were changed to INCLUDE Farm Labor, mechanical cotton pickers, etc., etc., appeared on the market. So the result was NOT an increase in income to farm workers. It was instead Unemployment, since the machines did a better job for less money. And you could just park them in the barn over the winter at no cost. (Note that this was also the HUGE advantage of automobiles and trucks: you just PARK the machines in the garage; you don't have to FEED them when they're not being used. The US Army shipped an entire Horse Cavalry Division to North Africa in 1942. The Army QUICKLY discovered that a motorized division was MUCH cheaper to operate, especially in an area where fodder for thousands of horses was hard to get. I believe 2nd CAV Division (Negro) was converted into transportation battalions. I watched a fascinating film about the modern method to grow tomatoes. 1) start with a HUGE hot house with GIANT hydroponic tanks, 2) "plant" new tomato plants in floats so their roots can soak in the tanks whilst the floats are mechanically circulated down the giant tanks, 3) take the developed tomato plants and hang their vines on hooks suspended from the ceiling, 4) every week or so harvest a batch of tomatoes and haul the next batch of green tomatoes up higher into the light, 5) when the original tomato plant stops producing green tomatoes, pull the plant out of the tank and throw it in the mulch bin. AI-controlled machines of course monitor and correct the lighting, the air temperature, the water temperature and chemical balance. Like the floor sweeper, humans are only used to do the bits too manually complicated to replace with machines. vmahuna
FF, I think we need to ponder mind vs brain and the question of conscious, responsible, rational freedom. In addition, we face the question of judgement informed by wisdom that goes beyond what we can articulate. Such judgement in complex, confusing and often one-off unique situations, is the mark of the expert practitioner. That is the context in which it is dubious that a programmed response is adequate. KF kairosfocus
Here is the biggest difference between current AI (deep learning) and the brain. Unlike the brain, a machine cannot detect a pattern or object unless it has a prior representation of it in memory. This is called representationalism. It is the current AI paradigm. The brain, by contrast, can instantly see a complex object it has never seen before without any learning. We would all be dead if our brains needed a prior representation in order to see new object. Non-representationalism is what AI researchers are focusing on, right? Not at all. They not only have no idea how the brain does it, they don't even care. They continue to hype deep learning as the path toward human-level AI because they are making lots of money doing it. FourFaces

Leave a Reply