Businesses beware: a new type of consumer is about to be unleashed, and they are armed with AI

There’s a brilliant passage (one of many) in the book “Requiem for the American Dream” where in fewer than 150 words, linguist and social critic Noam Chomsky pulls back the veil on what he describes as the century-old science of “manufacturing consumers”: 

If you’ve ever taken an economics course you know that markets are supposed to be based on informed consumers making rational choices. I don’t have to tell you, that’s not what’s done. If advertisers lived by market principles then some enterprise, say, General Motors, would put out a brief announcement of their product along with comments by Consumer Reports magazine so you could make a judgement about it.

That’s not what an ad for a car is—an ad for a car is a football hero, an actress, the car doing some crazy thing like going up a mountain or something. If you’ve ever turned on your television set, you know that hundreds of millions of dollars are spent to try to create uninformed consumers who will make irrational choices—that’s what advertising is.

Here Chomsky picks on television, but as anyone knows, when it comes time to seek out unbiased information on the internet, the unreality that he’s referring to is amplified 10 fold. Imagine that your TV now knows your location, your profession, your age, your gender, who your friends are, what your budget is (at least approximately), what other cars you’ve looked at. As Yuval Noah Harari has said, imagine ‘an algorithm that knows you better than you know yourself.’ 

That’s what consumers are up against when they set out to use the internet to make rational choices about what they do with their money. In some cases, like buying a car, a house, an insurance policy or a professional service (like financial consulting), these are choices that could have life-changing implications. 

Well, help could be on the way, because AI is making it possible for consumers to cut through the layers of deception and manipulation on which businesses have relied for decades, and such tools could allow consumers to make smarter, more rational decisions based on reliable, raw data, rather than insidious advertising techniques.


The last time I shopped for a car was just over three years ago, in early 2020. My wife and I had always done fine with just one, but when I was hired by the Sapienza University of Rome to teach a few English and writing courses, the commute made it impossible for us to get by without a second.

When the time came to begin my ‘customer journey’ (as marketers like to call it), I did what the majority of 21st-century consumers would do in my position: I opened my laptop and began to research my options online. Nothing I hadn’t done a million times before …

Needless to say, what ensued was about as far from a ‘rational’ process as one can imagine. I won’t bore readers with the details. Anyone who’s been in my position knows well enough what I’m talking about: wading through scores of online auto reviews, a labyrinth of rankings, lists and tables where objectivity is scarce, copy-and-paste rampant, and every detail so arbitrarily redefined that meaning seems to evaporate into thin air; dealerships masquerading exorbitant fees as ‘transparent pricing’, beguiling prospective customers with flashy landing pages, videos, and a seemingly endless list of add-ons and purchase options.

The worst part, in my mind, is the way in which your attention is constantly (and deliberately) interrupted. Every click sends you spiraling back into uncertainty, because you’ve just realized that you’ll have to pay extra for the color you wanted, or to have the eco-friendly LED headlights. So now you’re having to spend more to get the product that you originally thought you’d be getting. But that means that for the same money, you could have another, different model that you’d originally written off because it was too expensive, and the process starts over. 

Inevitably the figure that you’d originally imagined to be a reasonable amount of money to spend on a car begins to rise, which is just what the dealers and manufacturers want to happen. 

In the end, for better or for worse, I was spared from having to make a decision. Covid hit and all my courses for the semester were moved online.

But last year, circumstances put us in another impossible position: the city of Rome is rolling out a plan to ban all diesel vehicles from circulating within the city limits. Our old car will be unusable. As much as I support phasing out diesel and gas, my heart sunk at the thought of once again wading through the minefield that is the online automotive marketplace, at the sickening succession of spiels that I would have to endure at the hands of slick salespeople, at having to navigate the kafkaesque language of leases, financing and insurance options. 

And like any good consumer, I’d just play along, telling myself that somehow this was just how things are supposed to be.

Except, as Chomsky so duly points out, this is not how things are supposed to be. 


So here I am, earlier this year, like an innocent man in a kangaroo court, about to embark on yet another Google search horror show

I open my laptop and sigh. I pretty much already know how it will all play out, and I’ll be right,, because in a few weeks I’d be exhausted, ready to change jobs, move to the countryside and live on a farm. Within minutes Google knew that I was on the new car hunt and I started getting served ads everywhere — in my inbox, my social media feeds, the apps that I use. I’d gotten emails and even phone calls. And the whole time I’d been wracked with self-doubt. After all, I’ve got bills to pay, a family to think of, my son’s college tuition to save for. I didn’t want to screw this up, but it had also become clear that there was really no way that I could not screw this up. 

Because let’s face it, in the modern world, the consumer is not in the driver seat (no pun intended).

So what do I do? I start thinking that maybe the best option would be to kick the can down the road and opt for a lease.

Again, which is pretty much exactly what the industry wants to happen — to paralyse you with indecision, and then sell you your dream car with a leasing option that, on the face of it, appears to be the “prudent” choice. Sure, just put down a small deposit and figure it all out (a few thousand dollars in finance charges) later. 

Well played. Well played. 

So that’s where I was in early January. And then, just when I’m about to sign off on a lease, a glimmer of hope arrives.  ‘Hey, why not ask ChatGPT?’ a friend suggests. And I figure, I’ll give it one more shot. I mean, why not? I’d used ChatGPT for academic research (granted with mixed results). Heck, I’d even used it to make bread (a better experience). Why not get consumer advice? 

What followed was at least in the direction of what Chomsky had in mind when he talked about “informed consumers making rational decisions”. 

No images or video, no landing pages, no sexy fonts, no strategically placed ads to tug at the senses, no phone calls or grifters pulling you further down the rabbit hole. Just clear, plain language. 

A sigh, this time of relief, because now I’m starting to feel like I’m in control, because 2023 is not 2020. Because 2023 is the year that web 2.0 was blown to smithereens.

Because 2023 is the year that consumers were given the tools to bite back.   

Enter Large Language Models  

In the digital era, consumers face an insidious paradox: we are simultaneously overloaded with information and starved of meaningful insights. Every search spawns a torrent of results, with each website, blog, forum, and review vying for our attention. This supposed ‘wealth’ of data is, more often than not, a disorienting maze of contradictions and vested interests. And hidden within this labyrinth are the dark spectres of targeted ads, tracking cookies, and algorithmic bias, all working to subtly manipulate our perceptions and decisions. It’s less about empowering the consumer with knowledge, and more about funneling them towards the highest bidder.

The point is, our model for consumer information is broken. 

If designed and used effectively LLMs and AI in general could be at least part of a solution (and if AI companies are smart, they’ll see the tremendous demand that’s out there for such solutions). In this case, when faced with the ordeal of purchasing a new vehicle, ChatGPT turned out to be a surprisingly valuable ally, providing relatively clear and objective suggestions based on my criteria. The responses were helpful, the flow of the conversation was intuitive, and, equally as important, the entire process was free of distractions, allowing me to make a lucid decision without “marketing fatigue”.

The AI did a decent job interpreting the cryptic jargon with which the auto industry bewilders its customers. MSRP, invoice price, dealer holdback, destination charges – using the right prompts, you can get these terms explained in a pretty straightforward, jargon-free manner that most novices could comprehend. Something like having a personal translator bridging the gap between industry speak and consumer understanding.

I was able to carry out a thorough comparison of various car models that fit my requirements. As we know, ChatGPT’s training data ends in 2021, so at the time I had to copy and paste some data from the manufacturer’s website (today you can just use the Bing plugin to scrape data off the web). Customer reviews and reliability reports were dissected, safety ratings factored in. The potential future costs of repairs and maintenance were also taken into account. 

The final list of options was rational, not overwhelming.

It was, perhaps, most useful when it came time to explain the nitty-gritty of financing and insurance options. I actually came away with a better understanding of the nuances of car leases, the pros and cons of which most full-time economists might find it difficult to lay out, and in Italian no less (which I speak fluently, but isn’t my native tongue). Tools such as the Legalese Decoder promise to take such services even further. 

This was me in control, me making the decisions. This was me guiding the carriage, and not the other way around. This was at least closer to the ideal of an informed consumer making a rational decision than anything that a web 2.0 “customer journey” would have looked like. 


The rise of LLMs and AI-powered consumer tools heralds the dawn of a new era. A world where consumers are not just informed but empowered. We’re heading towards a future where the convoluted layers of marketing speak and deliberate obfuscation are stripped away, replaced by clarity, understanding, and the power of choice.

This change won’t just revolutionize how we shop, it will fundamentally shift the balance of power. Indeed, it may even convince us to shop less (how interesting that would be). No longer can businesses rely on a model that seeks to confuse, misdirect and manipulate. The informed consumer is here, and they are demanding transparency, accountability, and respect.

So, to all the businesses and brands out there: a new type of consumer is about to be unleashed, one armed with AI. The question is, are you ready?
In future posts, I’ll explore how the semantic web empowers consumers, and how brands must adapt in this changing landscape, or risk falling behind. Most importantly, I’ll be providing practical advice that brands can use to not just to stay relevant, but to excel in the era of AI and large language models by reigning in deceptive marketing practices and focusing, foremost, on transparency, quality and respect for consumers. I’ll also be laying out Semantic SEO and LLSO (large language search optimization) best practices.