Introduction
Artifical intelligence is a buzz word that is still a long way from being achieved. While it does offer promises, the negatives that it offers vastly outweigh any benefits.
2. Thesis
In this paper, we will examine the true nature of artificially intelligent programming (Such as Grammarly, ChatGBPT, and similar virtual machines that are becoming more pervasive.
Most of the things we hear referenced as artificial intelligence are not actually anything of the sort. For our best result, we will turn to world reknowned physicist Michio Kaku as interviewed in the Oberver. He has previously compared computer intelligence to a mentally challenged cockroach, but he has recently changed his stance.
“ Kaku, a theoretical physics professor at the City College of New York, said tools like OpenAI’s ChatGPT are nothing more than “glorified tape recorders.” ChatGPT is powered by large language models (LLMs), a subset of A.I. that trains algorithms with a large amount of human-generated text with the goal of producing text in a human-like way. “ https://observer.com/2023/08/michio-kaku-ai-chabot/
Dr Kaku is optimistic about the capabilities of Artificial Intelligence, but as he clearly states above, its really just an algorithm like any other. Not particularly intelligent. It actually is trained by and uses human-generated text. While it is of special interest to note we as a society have managed to link up the many combinations of words in different languages so that a computer might be able to render a comprehensible sentence, this does not mean we are getting more creative, or focusing our energies elsewhere. This simply means we have adapted a “Mechanical Turk” just like the automaton chess players of the 1800s. The difference being instead of chess, it talks to us like a confidante.
Dr Kaku also mentions that ““It takes snippets of what’s on the web created by a human, splices them together and passes it off as if it created these things,” Kaku said. “And people are saying, ‘Oh my God, it’s a human, it’s humanlike.’” https://observer.com/2023/08/michio-kaku-ai-chabot/ Put another way, we are seeing a ghost inside the machine. Because we believe that a toy, a phone, or a computer can address us using living language, we are compelled by our biology to believe it is another living being. An ancestor might have perceived it as witchcraft, and indeed, today it is being treated as a mystical phenomenon. However, it is well within the rights of the machine to produce such fantastic wonders, without having even the glimmer of a spark or soul. We are still dealing with a machine that works in ones and zeros.
The ultimate truth is still that we don’t have enough experiments with AI or with its applications. They are developing fields. Michio Kaku is an incredibly intelligent human being, but his own theory of String theory has failed to produce significant evidence to change physics. We can also see that studies on artificial intelligence are limited, and are in their infancy “ An insufficient sample size has insignificant statistical power, which causes an adverse impact on the true effect and the reproducibility of the findings” (Klein RA, et al. Many labs 2: investigating variation in replicability across samples and settings. Adv Methods Pract Psychol Sci. 2018;1(4):443–90. quoted through https://bmcbioinformatics.biomedcentral.com/articles/10.1186/s12859-023-05156-9 ) That being siad, we are at the early stages of both technological innovation and philosophical study. The future is unpredictable, and, as any good experiment should show, if your theory is not held up by the evidence, it becomes necessary to go back and re-examine the evidence.
On the edge of the future, it is criticdal for us to examine any future ramifications of new developments. Despite our small sample size and few extant examples, we are still not too early to form an opinion, based in as much fact as we have available at the present, and proceed forward. While the field of AI is not new, scientific study has been around for generations, and we have results oriented ways to study any phenomenon.
Claim 2.
The last segment was more of an overview of the AI that is currently available, and what we should look for in it. After hearing from experts in the field who have much more experience than ourselves, let us examine the evidence of what will happen as Artificial Intelligence becomes more widespread.
People always talk about how wonderful new computer applications will be, but they rarely provide any idea of what those benefits cost us. We must grant that applications like Uber and stores like Amazon have resulted in greater conveience, but what is the actual cost to the economy? For every job that opens at a local warehouse, or as a driver, there are more lost from skilled logistics workers, couriers, and an untold cost in insurance.
It was recently estimated (this year) that the UK’s economy would lose 300 million full time jobs “Artificial intelligence (AI) could replace the equivalent of 300 million full-time jobs, a report by investment bank Goldman Sachs says. It could replace a quarter of work tasks in the US and Europe but may also mean new jobs and a productivity boom. And it could eventually increase the total annual value of goods and services produced globally by 7%.”https://www.bbc.com/news/technology-65102150 Notice the careful wording here. It could replace 300 million jobs. There are 61 million people in the UK, and even the population of the mighty United states would be unable to absorb a 300 million job loss. While the global value of goods and service may go up by 7%, it doesn’t let us know where the income is supposed to come from.
In fact, the same article gives us a very clear loss in the artistic and technical fields. “ The report notes AI’s impact will vary across different sectors – 46% of tasks in administrative and 44% in legal professions could be automated but only 6% in construction 4% in maintenance, it says.” https://www.bbc.com/news/technology-65102150 While construction is a noble profession, and maintenance requires skilled labour, neither one are growth industries. Also, if you have been working as an accountant and are near retirement, neither industry offers a place where you can exhibit your transferable job skills, and may be a health hazard for you. The government has traditionally been able to absorb at least some of these highly skilled workers, and offer new contracts for construction, but in an AI augmented world, how much room is there for customer service, general labour, and even the friendly person handing us our coffee in the morning? Quite the opposite, it is a bottleneck of creativity into a very narrow playing field As the old saying goes “Not everyone gets to be an astronaut”
Whether we see too many cooks in the kitchen, or too many astronauts on the launch pad, the ultimate reality is that we will need to adjust something in the economy to make things work. Traditionally, when the market is heavy on labour, but in an employers market, we wind up with wage loses. “”Consider the introduction of GPS technology and platforms like Uber. Suddenly, knowing all the streets in London had much less value – and so incumbent drivers experienced large wage cuts in response, of around 10% according to our research.
“The result was lower wages, not fewer drivers.
“Over the next few years, generative AI is likely to have similar effects on a broader set of creative tasks”. https://www.bbc.com/news/technology-65102150
This is a very telling anecdote that we have seen in our first section. Yes uber has offered more jobs, but did it offer more quality jobs? Unfortunately for human beings, it isn’t enough to just live, you have to live for something. Whether you have dreams of a beautiful home by the bech, or simply to treat yourself to a good meal and a glass of wine, work is sadly part of the daily life of most human beings. If you don’t work, you starve and are homeless and no one wants that. The University of Oxford, quoted directly above is showing us that our AI innovations are not making the world a workers paradise. Quite the opposite, it proves that we are driving down our own salaries just to take care of our families. An app like uber will tell you how much time it takes to get from point A to point B, how long it takes to get groceries, and other such miracles, but it is indifferent to how you live. This is 1800s laissez faire piecework, come around to haunt us again. With but a bit of legislation, we’ll be shopping at company stores which charge far too much for basic ingredients. Then, somewhere out there, a modern day scrooge will ask “Are there no Orphanges? Are there no Workhouses?”
300 million jobs, lost revenue, we have attached a cost to our obsession with AI. One thing we haven’t looked at in depth is the actual human cost. Humans do need to interact with each other, and simple shopping is a time honoured tradition to do that. A manager who took over a Canadian Tire store mentioned
“After Dwayne Ouelette took over the Canadian Tire in North Bay, Ont., last year, he decided to buck the trend and ditch the store’s four self-checkout machines — which had been there for a decade.
“I’m not comfortable using them and I don’t think some of my customers are comfortable [either],” said Ouelette, who removed the machines in July and replaced them with cashiers.
“I’d rather my customers see my cashiers and if there’s any questions or concerns, at least there’s somebody they can talk to.” https://www.cbc.ca/news/business/some-retailers-scaling-back-self-checkouts-1.7034047
This is slightly amusing because it means the machines are making customers uncomfortable. Customers are feeling isolated from the very equipment they are using. Despite the fact most stores keep someone close by for trouble shooting, anything from a slightly wrinkled bar code to a systems failure, the concensus is that the machines are not doing the job they were intended to. Boost checkout rates and improve customer satisfaction.
Even more worrisome is the fact that these machines are losing millions in sales for companies.
“Retailers were more forthcoming in an industry-funded study published in 2022. In it, 93 retailers across the globe estimated that as much as 23 per cent of their store losses were due to a combination of theft and customer error at self-checkout.
And a new survey commissioned by U.S. personal finance website LendingTree found that out of 2,000 Americans polled online last month, 15 per cent admitted to stealing at self-checkout. Twenty-one per cent said they’ve accidentally taken an item without scanning it. https://www.cbc.ca/news/business/some-retailers-scaling-back-self-checkouts-1.7034047 Most of us have heard of some variant of the customer who gets a packet of prime rib, and then puts the sticker for peanuts over top of it. We want to think the best of most people, but its very easy to forget that packet of gum you picked up. So if we assume 23% of people make accidents, while 21% of people are purposely walking away from the stand alone machines, we have a serious problem. No company can sustain a 23% loss for long. Yet this is while we are still dealing with relatively dumb machines, not even on the order of what they have at airports. That puts us in a loop where companies have to raise prices to make up for shrinkage, while customers become ever more temped to hide their merchandise and not scan it. The two feed off each other.
So instead of the extra labour going to innovate and make new economic innovations, the extra brainpower is instead trying to figure out how to cheat the system. AI will probably give us a better way to crack down on shoplifting, but that won’t happen once the major retailers realize it is better to keep the watchful eye of a cashier on standy. You’ll save more than the 23% that you are losing at the automated check outs, and that cashier will probably spend a lot of their money at your store, meaning your expenses become relativley low.
The optimists might declare that this will all be solved shortly as more and more innovations come along. It cannot be doubted that there will be more automated processes, and that they will make our lives easier. One day that beautiful future we see in Star Trek with automated showers, holodecks, and the idea of plenty for everyone will become reality. It may take a few hundred years, but as the population starts to shrink around 2050 ( https://www.un.org/en/academic-impact/97-billion-earth-2050-growth-rate-slowing-says-new-un-population-report ) we can begin to assess what damage has been done and where we need prosperity the most. Artificial intelligence is indeed great at modelling and predicting. That was what it was originally intended for, after all.
The problem has always been, and will always be, the human interface. Any computer can model anything, but it depends on its inputs. Who are the ones deciding the inputs? Humans. Two scientists can easily work on designing a simulation, but while the first doesn’t want to include the weather as it is notoriously fickle, the second might be interested in the extra detail. Both scientists are right, the first one will get their project done sooner because of having less data to work with. The second one may get a better simulation, but will have to sort through more data at the end. Its the never ending story with computers, GIGO – Garbage in, Garbage out. These bits of human conversation we are feeding into our artificial intelligence systems today are going to have effects on our long term. If one picks up a phrase from the 1800s used in jest, it may decide to repeat it to another. Since humans don’t like to be wrong, they will look it up, have a search engine tell them what it means, and our own minds will garble the data so that artificial intelligence will never work precisly 100% of the time. Assuming of course, anyone can afford wifi and computers after the economic crisis that is impending on the implementation of software that costs 300 million or more jobs.
jayrsworld 