Category: Uncategorized

We taught machines to write like us. Now the question may be whether we still have anything to say.

The year is 2056.

My wife is trying to tell me something important. Six words in, an impeccably cheerful voice interrupts to recommend a marital-peace insurance policy.

I can’t complain. I’m on the free tier of marriage. Wife Plus, the ad-free version, is too expensive.

Besides, in fifteen minutes I have to log on for my monthly Reverse Turing Test. On the other side of the screen, an AI will determine whether I can still come up with an original thought, rather than simply remixing everything everyone else has already said.

I’m a little concerned. I just spent ten minutes trying to come up with a convincing answer to “How are you?”

But let’s go back to 2026, when Wife Plus was still just a meme on X.

And while we’re time traveling, let’s make one more stop.

In 1950, Alan Turing proposed a famous imitation game: could a machine converse in a way that made it difficult to tell apart from a person? The question was far subtler than the simplified version we tend to repeat today, but the image stuck.

I’m imagining a somewhat less scientific variation. Not a machine trying to pass for human, but a human being asked to demonstrate that they still have something of their own to say.

This isn’t a technical test. Nor is it a new version of a CAPTCHA, sometimes described as a “reverse Turing test.” It’s a provocation. And to me, the interesting question isn’t who typed the sentence. It’s who actually thought through what it means.

The Price of Words

Today we can produce a presentation, an article, or a business plan in seconds. Clean structure. Flawless grammar. Just the right tone. Often even enjoyable to read.

That’s progress. It saves time, makes previously expensive capabilities accessible, and helps people express ideas they might otherwise struggle to put into words.

But there’s a misunderstanding we may carry forward: making it easier to generate an answer doesn’t automatically make it easier to ask a good question.

Consider a business decision. We’re thinking about entering a new market. AI can help us compare scenarios, gather information, organize assumptions, and even challenge our enthusiasm.

Great. But which market do we choose? Which signals carry the most weight? What don’t we know? Which opportunity are we willing to walk away from?

The memo summarizing the decision might take a minute. A serious decision takes more than a minute—and more than a memo.

This Isn’t a Competition With Machines

The easy response would be to say that AI creates soulless content and human beings, by definition, don’t.

Convenient. Also not very credible.

I’ve read enough human-written material to know that a human byline doesn’t guarantee a thought. And I’ve seen how well-used AI can help sharpen an insight, catch an error, or pressure-test an argument.

AI-assisted writing can be rigorous. Writing produced entirely by a person can be pure filler. Authenticity of origin, quality of the result, and credibility of the claims are three different things.

“Soul” can sound vague when applied to content. I associate it with something less mysterious and more demanding: a recognizable point of view. The ability to notice what others missed, connect it to a real problem, and take responsibility for the conclusion.

A voice doesn’t come from an obligation to write every word by hand. It comes from the choices we make—including what we decide not to delegate.

The Part You Can’t Outsource

Picture an executive presenting a convincing strategy, developed with the help of sophisticated systems.

We ask why that particular course was chosen.

The executive can show charts, cite sources, summarize alternatives. Excellent.

Then we ask: “What would have to happen for you to change your mind?”

Now that’s an interesting question. Not because a machine couldn’t suggest an answer. It could. But because the answer ought to have consequences for the person giving it.

In business, this is where judgment, decision governance, and risk management come in. A plausible conclusion isn’t enough. You need to understand the assumptions behind it, what remains unverified, who is allowed to challenge it, and who is accountable when reality takes a different turn.

AI can reduce the work required to formulate a claim. It doesn’t erase the work required to verify that claim—or the consequences of standing behind it.

And perhaps the real danger isn’t that machines will end up speaking like us. It’s that we’ll get used to speaking before we’ve decided what we mean.

Back to the Future

Come to think of it, my imaginary Reverse Turing Test would make a terrible exam. There’s no infallible way to certify that a thought is original. And a sentence no one has ever uttered before can still be a spectacularly original piece of nonsense.

Maybe a better test is simpler: explain why you chose to say something, what supports it, and what might make you reconsider.

Artificial intelligence may free us from much of the labor of producing words. Not necessarily from the effort of having ideas.

In fifteen minutes I’ll have to answer the examiner.

I just hope my wife can finish her sentence before the next commercial.

From a mojito in Havana to Made in Italy in the age of artificial intelligence.

For a mojito, I would cross the street. At most.

Sonny Crockett crosses the sea.

In Miami Vice, Michael Mann’s 2006 film, there is a scene that has stayed with me. Crockett, played by Colin Farrell, asks Isabella, played by Gong Li, where she would like to go. She asks him what he likes to drink.

Mojitos.

Isabella knows a place that makes the best ones: La Bodeguita del Medio.

Crockett assumes she means the Florida Keys. A reasonable destination, considering they are already in Miami.

Isabella has other ideas.

Havana.

Before long, they are on a speedboat racing across the sea at a ridiculous speed, with a soundtrack that turns a rather questionable crossing into one of cinema’s most seductive journeys.

All for a cocktail.

Of course, it is not really about the cocktail. You can find rum, lime, sugar, and mint in Florida, too. There is no need to challenge geography, diplomacy, and a few border regulations to mix them in a glass.

It is something else.

How far are you willing to go to reach what you consider the authentic source of something?

How far?

The scene takes a question to cinematic extremes that, when you think about it, concerns a surprisingly important part of the economy.

The value of an address

A mojito at La Bodeguita is not necessarily worth more than another because it contains better ingredients.

It has a different kind of value because it is associated with a place, a history, a tradition. For Crockett and Isabella, getting there is part of the experience.

The point is not that La Bodeguita is necessarily where the mojito was invented. It is that the place represents a promise.

And promises, when they convince us, have economic value.

Made in Italy draws on this, too.

We do not just buy objects, food, or services. We buy skills, places, production traditions, and reputations built over time. And sometimes we simply buy the image we have formed of all those things.

The difference is less innocent than it seems.

What is the word “Italy” worth?

In 2021, a study published in Frontiers in Psychology analyzed how Italian origin and Italian Sounding influence the way people evaluate food products.

In the U.S. portion of the study, 237 consumers evaluated different packages of pasta.

The results on stated willingness to pay were interesting.

For a generic package, about $1.50.

For an Italian Sounding package, about $1.65.

For a Made in Italy package, about $2.

In other words, about 30% more than for the generic product.

A word of caution: these were evaluations collected in 2016, not actual purchases. They do not show that any Italian product can command a 30% premium. Nor that simply stating its origin is enough to obtain one.

But they suggest something important.

Perceived Italianness can help generate value. And some of that value comes through the product’s reputation.

The problem is that Italian Sounding can benefit from it, too.

An Italian name, evocative packaging, a familiar image. All can suggest a connection to Italy, even when the nature of that connection remains unclear.

Not every cultural reference is counterfeiting. And not every Italian-inspired product claims to have been manufactured in Italy.

But if Italianness creates value, it becomes important to distinguish what is being evoked from what can be documented.

When the story meets a machine

Until now, a substantial part of marketing has worked by building associations.

A Tuscan landscape. A family around a table. An artisan’s workshop. A few Italian words, preferably delivered with a convincing accent.

Nothing necessarily wrong with that. Stories help communicate value.

The trouble begins when the story becomes more precise than the available facts.

And today, a new intermediary may make that gap even more interesting: artificial intelligence.

Imagine an American consumer asking a digital assistant to find an extra virgin olive oil produced in Italy, compare a few olive mills, and explain what information supports the stated origin.

They are no longer simply looking for a brand that reminds them of Italy.

They are asking to connect a promise to verifiable information.

Of course, the assistant can get things wrong. It can confuse sources, repeat unsupported claims, or draw conclusions the documents do not support.

And it cannot replace a supply chain inspection.

But it can search and compare publicly available information, if that information exists and is understandable.

This is where the question changes.

It is not enough for a business to have an authentic story. We also need to understand whether that story can be pieced together accurately by someone who does not already know the business.

Human or machine.

The possible dividend of verifiability

Imagine two equally reputable producers, both with a real connection to Italy.

The first tells a generic story of family tradition and shows beautiful photographs of rolling hills.

The second tells the same story, but also makes clear where it designs, where it produces, where its raw materials come from, and which documents support its claims.

We have not demonstrated that the second will sell more.

And it would be rather naive to think an algorithm automatically rewards those who have their paperwork in order.

But the second has made something possible that the first has not yet done: distinguishing the narrative from its factual basis.

This could become a form of competitive advantage.

Not a new universal certification. Not a seal invented to reassure a chatbot.

Rather, the ability to make existing value accessible and understandable.

With one essential precaution: what is not publicly documented is not necessarily false. And what is documented does not automatically become true simply because an AI system can read it.

Verifiability is not a substitute for reality.

It is a way to let others examine it.

Back to Havana

There is one last detail about the Miami Vice scene that I find irresistible.

To recreate Havana, Michael Mann also filmed in Uruguay.

Of course, there is nothing wrong with that. Cinema has to convince us we are somewhere, not prove that the crew was actually there.

But it is a small reminder of how powerful a well-crafted representation can be.

We can recognize a place without ever having visited it. We can associate a story with a product without knowing exactly where it comes from.

And we can pay for that story.

The question, then, is not whether storytelling has value. It always has, and probably always will.

The question is what happens when the person making a choice starts asking for evidence, too.

Perhaps the real leap is not about taking every American consumer to visit Italian producers.

It is about making the reverse journey possible: bringing them not just the promise of authenticity, but also the information they need to assess it.

Sonny Crockett needed a speedboat.

In our case, perhaps we could start with the facts.