Who really wants AI…and why?

TE Creus

I’ve noticed that self-checkout is taking over more and more shops and supermarkets. They have existed for a while, of course, but more as an alternative to reduce lines, but now many shops have exclusively self-checkout machines.

Yet, self-checkout appears not to be very effective or convenient, neither for shops nor for customers. Many customers tend to dislike self-checkout, as evidenced by the fact that there are always lines for human cashiers but none for self-checkouts. They give you the perception of more speed, but it is just illusory. Professional human cashiers scan and move your products faster (especially in Germany, where they go so quickly, basically throwing the stuff at your face, that you can hardly keep up).

“Nobody likes self-checkout”, says an article at CNN, “Here’s why it’s everywhere”.

Basically, as it is typical in the “digital economy”, it is just another way of passing the work to the customer and making think he’s gaining something with the exchange. Now, it may work for some — and I guess it is good if you want to avoid human interactions with a cranky cashier, which sometimes has its benefits.

You’d think that this type of automation would reduce the work of human cashiers and therefore save money for the companies and therefore make products cheaper, but it is not so. First of all, the machines need constant maintenance. Even if companies reduce the number of cashiers, they need to hire more technicians, which are paid more. The machines cost a lot, too, and require programming. And are supermarket products becoming cheaper? I don’t think so, quite the opposite in fact…

But not even the reduction in the number of human cashiers is a given. When people have to do their own checkout with machines, there is always something that goes wrong, or some product that won’t be scanned, so many people constantly require assistance even when using the self-checkout. In the end, cashiers and supervisors actually end up having more work, instead of less.

Of course, self-checkout is more conducive to shoplifting — sometimes voluntary, sometimes accidental. There are items, such as bread, that have no barcodes, and it becomes more complicated to register them in. Even if most people are honest, some are not. Companies lose more money with that, too — according to that same CNN article, losses are about 77% higher than at stores without self-checkout.

Since it doesn’t save money for companies, nor makes the products cheaper or the experience better for customers, one has to wonder why most companies are moving to self-checkout anyway?

CNN’s and The Guardian‘s answer is that “most companies are doing it, so they feel that they have to do it too”, which seems a stupid explanation.

Unfazed by the relative failure of self-checkout, Big Tech is pushing for even more of it. Amazon, who owns Whole Foods and other physical stores, has introduced “smart carts”, where your products are scanned and weighed as soon as you put them in the cart, no checkout needed (your debit card or phone is charged automatically).

Other new versions of self-checkout include shops where each movement is tracked by AI cameras and motion sensors, registering each item you take from the shelf and billing you later. All you need to do is swipe a credit card or smartphone, at least until the new methods that allow you to pay by facial recognition are installed.

The same is true of customer service — most people dislike bots as costumer service agents and prefer talking to a person. Yet bots keep being used, more and more. I suppose they are cheaper than paying a person to answer the phone, but my suspicion is that it’s about something else entirely. And in fact an article gives it away when it says that bots can easily record, memorize and access all your information, in order to provide for a “better customer journey” (the new marketing buzzword is “journey” instead of “experience”).

Self-driving taxi cabs, the same thing — the companies working with that are not making money, and even if you reduce costs by removing the driver on one side, you increase it on the other side by having to hire more technicians and supervisors — or even just someone to clean the car, as a Roomba can’t do it.

Obviously, there is something else behind the hype. Like with everything related to AI, there is a huge push by Big Tech for making everything automated and AI-dependant, from cars to journalistic essays to art works. And the reason is “big data”. Getting all the information they can about you.

In “1984”, George Orwell predicted screens observing us 24/7. But now we have not just screens, but cameras, satellites, location tracking apps, facial identification apps, voice recording apps, AI. They will know what you eat, what you poop, how much money you make, how much you spend, what hereditary diseases you have, and what you did last summer.

In the end, what “Artificial Intelligence” is really about, is not “intelligence” in the sense of “being clever”, but “intelligence” in the meaning understood by the CIA — gathering “intelligence”.

Recording, remembering and accessing all kinds of information from everybody.

Soon, all objects are going to be spying on you.

And it will not be just governments and big corporations. A recent story about a mother who received a message about her daughter’s fake kidnapping using the girl’s cloned voice, as well as the emergency of very believable “deep fake” videos, show the huge boost for scamming and crime provided by these new technologies. If you thought spam mails and spam calls were a nightmare, get ready for the new AI-powered identity thefts and scams.

And, of course, there is also the issue of “sensitivity filters” being applied to AI. Right now, they are using mostly human “sensitivity readers” to rewrite novels from Roald Dahl to Ian Fleming to Agatha Christie — despite the growth of AI, which is supposedly “stealing all our jobs”, there is a growing number of absurd, unnecessary professions, such as “DEI consultant” or “sensitivity reader”. And yet, one would think that this is one of the rare cases in which AI could probably do a better job — how hard can it be to write a program to filter out “offensive” words and replace them with something else?

Not that I support any of that, of course. Besides the more general attempt to change the past (“who controls the past controls the present”, as per Orwell), I think what is really going on here, under the excuse of “not offending anyone”, is simply an attack on readers, and on reading in general. Not too many people read books these days, and if all books are just going to be rewritten into the same inane, boring, nondescript, corporate propagandistic jargon that is the norm everywhere else in the media, well, then what will be the point of reading, anyway? You might as well have all books written by AI bots — and read by bots, too, since such books won’t interest anyone else. Which I suspect is the idea. Murdering literature, and getting away with it.

The other day a friend showed me one of these new AI chat toys. You type in some words and the bot creates a silly story for you. The first attempt wasn’t very memorable, so to make the story a bit more exciting, in his second attempt, my friend typed in a few new random words, one of which was “killer”.

Instead of coming up with a story, the bot replied some boilerplate excuse that it is programmed for peace, understanding and tolerance and thus could not create a story that promotes or shows violence. (There goes the idea of using it to create the next best-selling mystery thriller.)

Meanwhile, on the sidebar, the same search engine that promotes the bot was showing all kinds of news about the latest daily school shooting, a stabbing in a kindergarten, and people being pushed by a psychotic homeless onto the subway tracks…

I thought it was very representative of the world we live in.

TE Creus is a writer and film-maker. He usually blogs at contrarium.substack.com and his movies are at contrarium.org.