HOW AI IS MAKING LIFE BETTER...FOR THE TOP 1%...

AI has a big impact on the world, with both actual useful uses and more dystopian ones. Read about how AI today is being used in more malicious ways than you might think to make the big bucks for huge corporations and their billionaire CEOs, and how it may affect you and others.

PIE CHART REPRESENTING COMPANY USE OF AI; GRAPHIC BY LILI NANDELL

AI here, AI there, AI everywhere. But what does it actually mean?

You’ve probably seen AI in the news, or in your feed. The next big technology that’s gonna change the world! Or whatever they say.

All these companies say that they’re putting AI in their... well, everything, and that that means it’s making their business better, enhancing their workflow, empowering customers, etc. What does this really mean? 

Well, let's start with the basics. You’ve likely heard of or used some AI program like Chat GPT. You give it a prompt or some other information, and it spits out image, text, video; it all seems pretty believable, especially if you only glance at it. AI programs take an input and spit out an output using neural networks. These put inputs through a network of nodes that change the number inputs and output data that can be decoded into media, movements for a digital creature, etc. 

This isn’t new, either. You might be surprised to find that neural networks and AI have been around for years! Some of the earlier breakthroughs in AI were around the 1960s and 70s, but all of this is actually based on 500-year-old work in early statistics .

AI HAS THE POWER; GRAPHIC BY SAMIRA KUNATE

So, neural networks can be very powerful tools in the way that they produce believable outputs that seem real or human, or ones that are just good at one particular task - but how do we get them to do that? It’s not like we just go through every one of the thousands of nodes and adjust them manually. Well, you train the model, which can be a difficult and very expensive process. The idea is simple: take a bunch of paired input and expected outputs, and train the model on that data. Now, this is pretty complicated, but to summarize: AI models can be trained through a variety of processes using training data. These processes, in different ways, get models to mimic their training data by adjusting their nodes in their neural networks. 

So, we now know the basics of how AI works and what it is. AI takes in inputs and spits out outputs, mimicking its training data (usually with some intentional added randomness). So how is AI being used in malicious ways by large corporations and others, and what can we do about it?

Automated, Emotionless Optimization

AI is automatic. A computer just runs some numbers and bravo, another episode in your totally epic series of AI cat brain-rot created and ready for you to upload to YouTube. Now, from the perspective of a higher-up corporation, this is insane. This is for the simple reason that AI can do both of the following: do tasks that are human-like enough while still being not a real person with ethics and a life, and optimize certain things, for example, predicting the budget of a person within a few seconds just by looking at a picture of them. This brings us to our main topic of conversation.

Personalized Prices

No longer restrained by the threat of human workers snitching and their potential inefficiencies caused by, ugh, ethics, corporations have recently been getting to work. By combining all the personal information they have of you and pictures or video from cameras or security cameras, they can run all of it through an AI program to decide what price they think you will be able to afford. You could walk around a grocery store (one that has implemented this tech) and see electronic screens change the price of an item right before your eyes. 

Now, personalized prices aren’t exactly new. Air travel companies have been playing around with personalized prices for a bit now, basing prices shown on their websites on availability and demand. 

These new personalized prices that are being experimented with in stores can be much more malicious, though. First of all, electronic screens can show ads instead of prices to increase revenue for a store. Second, prices can be placed in weird places on the screen and/or only some prices can be shown to make it confusing which price goes with each product. This could potentially get rushed, not-paying-attention, or novice shoppers to grab a product thinking that the price they saw on the screen went with that product, when it really didn’t. Finally, these prices are always changing. The price you see on the shelf probably won’t be what you see at checkout, and companies can abuse this to make the price at checkout much higher than at the shelf. This means that you might see a good looking price at the shelf and grab an item, but then have to pay a much higher price at checkout. You might be less likely to return the item because going back might look weird and would be an inconvenience. Stores might even make it harder or against the rules to return an item to try and guarantee your purchase. Plus, you’ll be constantly watched and tracked throughout a store in order to update the electronic price tags around you, for you. That way we can all feel uncomfortable when we’re out trying to get a jug of milk!

Now, because the AI that will be used to determine the budget of somebody will be trained off of real data (who could’ve guessed), it will discriminate. Skin color? Yep. Gender? Yep. Mother carrying their two children with them to the store? Probably lower income with father out working and family unable to afford daycare. The children will need food and might want toys, and that 2 year old is going to need diapers. The clothing the mother is wearing also hints at their income status.

So, personalized prices can change the price based off of what the AI thinks you can afford, making (sometimes creepily accurate) sweeping generalizations just from a few pictures of you. So, how might personalized prices affect you? Overall, this article has been quite negative towards personalized prices, but if companies decide to take a certain approach, personalized prices might have some benefits. If a company knows that a customer is willing to buy any brand they might lower their price to be competitive. In this scenario, personalized prices have some benefit. However, companies are companies, and companies are greedy. 

As they’ve shown in the past, companies will likely use personalized pricing to extract as much money as possible from everyone, and manage loyalty not through prices but through other means. Personalized pricing also discriminates, undermines customer confidence, and makes it impossible to know what a price will be until you’re in the store. Personalized pricing also encourages more personal data collection, making privacy less private.

It can be pretty dystopian what these systems can do, and because they’re all just one big automated thing; a few numbers can be changed to change how the whole thing works. Easy as that. Imagine this scenario: A grocery store company wants a bill to be passed that will increase their profit margins by allowing them to decrease safety for staff and customers. 

To get the bill to be passed, they want to get a mayor elected. The store management can change some numbers in the system and add some advertisements to make it appear that the current mayor (who is running for reelection and won’t pass the bill) is wrecking the economy and making prices unaffordable. Yet to ensure they still make money, they only show this to people who aren’t already going to vote for the mayor who will pass the bill. That way, they can sway the votes in their favor and get the bill passed.

You might think that personalized pricing will make life better for poor people, because they might be able to afford items with their low income. Yet, once you realize that as their condition gets better and they make more money, prices will just go up and the amount of things they can afford will not change at all. This might even spread to everyone, where even at a decently high income, anyone would only be able to afford a few items from the store to eat.

Summary and Opinion

AI is... a new thing. Like all new things, it’s new. It will do all its new thing stuff, before a new new thing comes along and takes its spot.

AI is being put in everything, likely to our detriment. While I think that AI has its legitimate and good uses, I do not think that it should be put in everything. 

Generative AI in particular harms trust and at times makes it nearly impossible to know if that image or that essay is made with effort by a real human, or by a machine that got a few words as its prompt. 

AI also makes it harder to be properly informed, because when websites are rewarded for clicks and not factual content, it’s more profitable to just AI generate articles or blogs to attract clicks instead of hiring someone to slowly make an actual piece of content. Personalized pricing will only further incentivise the collection of people’s personal information, and the swath of strategies companies could use with it to increase profits is quite large. 

Websites used for information:

1: https://en.wikipedia.org/wiki/Neural_network_(machine_learning)

2: https://insights.som.yale.edu/insights/the-perils-of-personalized-pricing