Gwangi`s thoughts on how AI will affect AS and AD

 (My comments are in italics).

Gwangi Kim (class of 2017) has found a very interesting video about artificial intelligence (AI).


Gwangi considers that our brains are nothing more than a very complicated machine. As AI develops, he argues that society’s rosy, yet blinded and romantic view of human creativity will soon be revealed as invalid.  He admits that creative jobs will be difficult to replace by AI  (while other office jobs, will definitely be replaceable) but it will eventually happen. Moreover, the proportion of jobs that use creativity in society is very small relative to other types in the workforce.

Therefore, Gwangi concludes that there will be excess supply in the long term. (This could also be phrased as excess aggregate supply  (AS) -a concept which 11th graders have not yet encountered – or excess capacity). Aggregate demand  (AD) in more economically developed countries (MEDCs) will not be enough to meet aggregate supply (which is exponentially increasing due to advanced technologies such as AI)  due to the substitution of labor by capital or by the cheaper labor force in other countries.

However, a possible and the only solution would be if an international government were to impose a tax on production and use the tax revenue to provide a basic income guarantee for everyone. Consumption would rise and this would lessen the excess supply problem, though probably not solve it.

Gwangi`s analysis is based on the fact that AI doesn`t demand goods as labour does. Humans and AI are similar in the fact that they both act according to basic instructions: DNA and the algorithm. However, the `goal` of DNA is the survival and reproduction of human beings; therefore, humans consume goods so that they can satisfy their biological demands. While the `goal` of the algorithm is to achieve the best performance in the task given; therefore, AI doesn`t demand any of the goods produced.

If Gwangi is right, this has serious implications for the future structure of economies and even, perhaps, for your own futures in the work-force. I have attempted to rebuff the gloomy predictions by suggesting that in the economically less developed countries (ELDCs) there are still many needs that have not yet been met. Demand will arise from these consumers so long as they can obtain the funds necessary (either from income generated by work or from redistribution, perhaps through the international tax suggested by Gwangi). I would also like to see “job-sharing” schemes arise, where two people share one job and each work for half the time. But I am not confident that such developments would avert the problem.


5 responses to “Gwangi`s thoughts on how AI will affect AS and AD”

  1. DevanshiRuparel says :

    Hey! I’ve recently created a blog where I post short articles ranging from happiness to DIYs, poems to facts about me. I’d truly appreciate it if you could check it out! 🙂

  2. Paul Elder says :

    Prior to watching the video, I was convinced of the things that the presenter assumed I was convinced of: white collar and creative jobs being more difficult to replace. But the video, with its evidence, was able to persuade me that the apocalyptic future where there would be no more jobs for humans is pretty realistic. Of course robots can do manual labor more accurately than us. Of course robots can process data more accurately than us. Plus all of that at a lower operational cost.

    My viewpoint is against this theory, so I’m going to try my best to gun it down. I am convinced that what the videos shows is not the whole story.

    I guess we’re assuming that all robots have low operational cost since they are to be developed to the point where they are very efficient with their electricity use.

    First of all, with no jobs, people have little income, if any, and would thus not be able to or will be unwilling to consume much. So what will happen to all the goods that are produced by the low cost robotic labor? Would that not result in a welfare loss? Since the cost of the robotic labor is so low, it would not incur a loss for the firms so they won’t care about reducing the wasted resources. Although no sales would hit the salaries of the higher-ups and lots of firms would go bankrupt and people would go back to manual labor.

    Another issue is the developmental costs of the robots. It probably costs a lot to be able to make robots that can work on par with intelligent and creative human jobs, so greedy (or sane) developers would put a high royalty on their robots, which would increase the costs of using a robot. This could probably save the human job market.

    Another problem with robots taking over the workforce is that they don’t care about the environment. I suppose you could try to teach them, since the robots could gather environmental information from a network, but ultimately it won’t affect their performance of their “life quality”. The wasted resources won’t concern them, and humans will kind of just die out.

    Another problem (I was going to make it an assumption but it helps my case so I’m making it a point) is that these robots are reliant on electricity. Of course the mobile robots will have batteries or generators and rescue robots are designed to survive a long time without needing to charge. However, other types of automation will require a stable power source. This means a highly defended power plant and highly defended power transportation (wires). I guess Google’s datacenters have a whole ton of backup generators so they have no problem, but remember a big earthquake 5 years ago that took out the power for half of Japan? Events like that could incapacitate whole workforces and incur even more costs than human workers taking a sick day off (theoretically).

    The biggest problem that I see with automation is that in order to work, the robots need a network for computing and information gathering. This leaves it vulnerable to cracking (hacking). Even right now, 85% of the internet (webservices) does not use simple security and encryption, and a surprisingly large amount don’t even do the simplest of security and they save passwords in their database in plain text. Even with the rise of the Internet of Things (connecting everything to the internet), most people don’t have even simple security, and you can even search the web for “vulnerable IoT webcams” and watch other people’s “security” cameras.

    Even if the security was beefed up, you would still have government and intelligence agencies (FBI, NSA) that would try to force firms to open backdoors into their robots, but believe it or not, putting the key under the doormat allows anybody to break in, likely making the jobs of crackers easier. Once they get in, they can do lots of damage, and in fact it has been shown that self-driving cars can be hacked and cause an accident and all traces of the hacking are erased, resulting in the human driver potentially getting in trouble. What’s worse is that the multi-purpose robots are multi-purpose, so guess what happens if a cracker manages to “teach” a robot face recognition and how to kill… It’ll probably be a lot worse that a credit card database heist.

    Also (little point) I think government work will still be human.

    So robots would take over all human jobs, including robot maintenance and the production of power, human necessities, human recreation, and more robots. That leaves practically everybody unemployed, requiring welfare payments from the government who probably doesn’t have much either thanks to low tax revenue. The only firms that will be make a profit are those producing necessities. This will essentially halt the flow of money, and humans would probably revert to self-sustained farming to produce their own food (that they can actually afford) and that would probably create informal markets that only have humans.

    On the other hand, exports to ELDCs would be pretty well off thanks to the (presumably) low prices which would provide income to the exporting firms… which barely have any humans in them. So I guess life is good for ELDC people and bad for MEDC people.

    And then after a while (or skilled speculators) in MEDCs would try to bring down the robotic workforce electronically. The robots (probably very intelligent by now) could probably fight quite well. the outcome is unknown, but if the humans succeed then they could get in trouble, though firms may start to see the benefit of human labor and start employing humans again.

    I doubt humans will bring about their own doom. That scenario is the worst case. Mmmm… this is a complicated situation with lots of consequences to consider. I realize I kinda just splattered my ideas here, but you can see what I think about this. I doubt all human jobs will be taken over, at least not for a long while.

    • econyamada says :

      Thank you, Paul, for your well-considered and forcefully argued thoughts. Although I am not sure whether I should be “thanking” you, since your vision of the future is even grimmer and more apocalyptic than Gwangi’s.

      At the end of your comment, you seem slightly more optimistic. However, it does seem that steelworkers in Britain are already feeling some of the consequences of the structural changes brought about by technology. It occurs to me that the decision by the Indian company, Tata, to sell off their loss-making steel plant in Britain (which they only bought–an example of FDI from India into Britain-in 2007) is due to those structural changes. Global demand for steel is no longer sufficient to meet global supply. On the demand side, miniaturization and computerization has meant that less steel is needed, while, on the supply side, new techniques enable steel to be produced in vast quantities at lower average costs.

      Despite all these structural changes, I remain hopeful that 11th and 12th grade students can look forward to interesting, satisfying and challenging careers. But you will need to be flexible and creative.

      • Paul Elder says :

        You’re welcome!

        I realized that my organization wasn’t very coherent and was just a stream of thoughts.

        I see that the steel industry is experiencing the same effects, but for a different reason.

        Now that I think about the AI issue again, I doubt that robots will take over the workforce, at least not for a long, long while, since even though the operation costs are very low (energy and maintenance costs only), the cost to switch the labor force would be a lot, and it would be for the many thousands? millions? of firms in each huge developed economy in order to cause a global job apocalypse. In the meantime, the firms that do employ a robotic workforce would lower production costs, though it is debatable whether they will lower prices for consumers to enjoy or keep the prices high for a welfare loss but increased profits (model of monopoly, with super-low MC, means that the profit-maximizing quantity would be more than a higher MC. Perhaps the price may decrease.) However, if the price of everything falls as production costs are lowered, could hyperdisinflation or hyperdeflation occur? I suppose this would follow Gwangi’s model as well, since the AS curve would shift so much right, causing more and more disinflation.

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

%d bloggers like this: