THE COURAGE PREMIUM

Perfection has lost its signaling power.

For most of modern consumer culture, polish carried information. If something looked beautifully made, sounded carefully written or felt painstakingly designed, you could reasonably infer that someone had invested time, expertise and resources in making it that way.

The finished object contained evidence of the labor behind it. AI has broken that relationship.

In a world where perfection is a prompt away, people are quickly recalibrating how they understand the things around them. The familiar signals of quality no longer carry the same weight, forcing us to search for new forms of validation, intention and trust that help us decide what is worth believing, valuing and buying.

EFFORT WAS PART OF THE VALUE

Long before generative AI, researchers had already identified something called the effort heuristic: when people believe more effort went into making something, they tend to assign it greater quality and value.

In a series of experiments published in 2004, participants rated art, poetry and craftsmanship more highly when they believed the work required greater time and effort to produce. The effect was particularly strong when quality itself was difficult to judge.

The result makes intuitive sense – effort is information. When value is difficult to discern we look for signals. Time invested suggests seriousness. Difficulty suggests skill. Craft suggests care. The cost of making something becomes part of the evidence we use to decide what it is worth.

Generative AI hasn’t eliminated that instinct. It has simply made the old signals unreliable.

A 2026 study revisiting the effort heuristic in AI-generated art found something remarkable. Participants could appreciate AI art aesthetically at levels comparable to human-created work, but they were significantly less willing to pay for it. The researchers found that the penalty was driven primarily not by some mystical belief in the human soul, but by perceived lack of effort.

When researchers made the human involvement behind the AI work more visible, its perceived value increased. The art didn’t change. The perceived investment behind it did.

That distinction is quickly becoming one of the defining tensions of the AI era.

A CRISIS OF EVIDENCE

The trust problem goes beyond whether consumers like AI.

Gartner found this year that 57% of U.S. consumers say the prevalence of AI-generated content has made them less trusting of brand messaging overall.

At the same time, 43% say they are relying more heavily on real people for shopping information and recommendations because of the growth of AI-generated content.

That’s not necessarily an ideological rejection of artificial intelligence. It may be something much more basic – people are losing confidence in their ability to tell what represents genuine investment.

A separate Gartner survey found that 68% of consumers frequently wonder whether the content and information they encounter is real.

The skepticism is becoming ambient. MIT researchers demonstrated how quickly that uncertainty can spread. In an experiment involving realistic AI-generated video, people who first encountered synthetic footage subsequently became less confident that authentic human videos were real, even though the AI material had been clearly disclosed.

The synthetic content didn’t just create suspicion of the AI generated content. It contaminated trust in everything around it.

That suspicion creates an authenticity tax for everyone, and it helps explain why simply making AI output better does not necessarily solve the problem.

Researchers at Berkeley found that consumers often could not reliably distinguish AI-generated advertising from human-created advertising. Yet ads believed to be human-made received higher evaluations across every effectiveness measure, regardless of who had actually created them. The highest-rated ads in the study were, ironically enough, AI-generated ads that people mistakenly thought humans had made.

The machine was capable of producing the output. What audiences valued was the belief that someone had meaningfully participated in creating it.

The issue was not quality alone.

It was provenance.

PERFECTION IS EVERYWHERE

This is happening at exactly the same moment organizations have become extraordinarily good at optimization. We have spent decades refining our ability to identify what works and reproduce it more efficiently. AI has compressed that process dramatically, making competent analysis and competent execution available at a scale we have never experienced before.

This is an extraordinary achievement. It’s also producing a strange side effect.

The hallmarks of quality are no longer reliable. When everyone can generate polished work, polish becomes less distinctive. When everyone can identify the proven answer, the proven answer becomes less valuable. When everyone can produce something technically excellent, technical excellence alone stops telling us very much.

Ipsos and Syracuse University’s Newhouse School recently tested human-created advertising against ads produced entirely by AI from comparable strategic briefs. Most consumers could not reliably identify which was which. However, the human-created work produced 14% greater short-term effectiveness and 17% greater long-term brand effectiveness. It was more likely to be perceived as distinctive, imaginative and worth talking about.

The AI work was already remarkably capable of making advertising that looked like advertising. And that may be precisely the problem. As perfection becomes ubiquitous, that ubiquity destroys its signaling power.

PROOF THAT SOMEONE WAS THERE

Markets have a funny habit of making whatever becomes scarce more valuable.

We are already beginning to see brands react.

Prada’s 2026 holiday campaign, Drawing Prada: Animated by Hand, combines live action with traditional hand-drawn animation led by veteran animator John Pomeroy. Prada goes unusually far in describing how the campaign was made, emphasizing that the illustrations were executed by hand and positioning the work explicitly as a celebration of craft and process.

The process is not hidden behind the finished product. The process is part of the product.

That distinction matters.

The audience isn’t simply seeing an illustration, it’s seeing evidence that someone drew it.

And Prada is hardly alone. The Wall Street Journal reported this week that advertisers are increasingly turning toward labor-intensive techniques such as stop-motion, practical effects and hand-drawn animation as AI-generated work proliferates. Production studios are seeing growing demand for approaches where the effort behind the work can be made visible.

TBWA made the same idea foundational to its own identity earlier this year. Its global rebrand deliberately embraced hand-drawn marks and imperfection, describing the design system as a bet on humanity in the age of AI.

These choices aren’t compelling because hand-drawn is inherently superior to image generation. Nor should every brand now commission a stop-motion mascot, lest we replace one exhausting convention with another.

What matters is what these choices communicate.

Someone was there. Someone made a decision, invested time, and cared enough not to choose the shortest path simply because it existed.

In an environment where the appearance of effort has become almost effortless, the evidence of actual effort becomes meaningful again.

THE RETURN OF COST SIGNALS

There is an old idea in economics and biology called cost signaling – a signal becomes credible partly because producing it costs something. The peacock’s tail is persuasive because it’s inconvenient. Credentials have value because acquiring them requires time. A warranty means something because honoring it exposes the company to financial risk.

The cost makes the signal harder to fake.

Something similar is now happening with human effort. For decades, technological progress has largely meant removing friction from production. But AI introduces an intriguing reversal. When efficiency becomes available to everyone, choosing where not to optimize can itself communicate value.

A brand that invests weeks in something it could technically generate overnight is not merely choosing a production method. It is making a statement about importance. This mattered enough to spend something on – not just money, but time, attention, resources, and craft.

The investment itself becomes part of the message.

THIS IS NOT AN ARGUMENT ABOUT AI

It would be easy to turn this into an argument for retreat. That would miss the point.

AI is already enormously valuable precisely because it removes unnecessary effort. There is little virtue in spending six hours doing something a machine can do better in six minutes simply to demonstrate that your suffering was artisanal.

The question is not whether effort is good, but where effort carries meaning. In the countless tasks where nobody cares whether a human touched them, efficiency should win. But there are other moments where what we are buying is inseparable from judgment, taste, authorship, expertise or trust. In those moments, optimizing away every trace of human involvement may actually optimize away part of the value.

This is the distinction organizations need to understand. Not “Where can we use AI?” – we already know the answer is almost everywhere. The harder question is, “Where does human investment itself contribute to what people are buying and buying into?”

That answer will be different for every organization. Finding it requires something optimization alone cannot provide.

THE COURAGE PREMIUM

For years, organizations have been taught that better intelligence reduces risk. The underlying promise is seductive: if we can gather enough information, uncertainty eventually disappears.

The reality is that it does not.

At some point the evidence ends and somebody still has to decide. And the more powerful our analytical tools become, the easier it is to mistake what they can prove for the boundary of what is worth doing.

The economically rational decision may be to automate. The benchmark may recommend the familiar solution. The data may tell us what reliably performs. And still, there may be reasons to choose differently, not because we reject the intelligence, but because we understand what the intelligence cannot measure.

Trust.

Meaning.

Surprise.

The feeling that someone cared.

This is where courage becomes a competitive advantage.

Courage begins with intelligence, but it also requires understanding the evidence well enough to recognize where it ends. Once the data proves that an opportunity exists, everyone can see it. And once everyone can see it, differentiation has already begun to disappear.

The valuable territory exists in the space between knowing enough to form a conviction and waiting until there is enough evidence for everyone else to share it.

That distance is uncomfortable. It’s also where distinction lives.

EVIDENCE OF CARE

AI has made something possible much faster than many anticipated – it has made perfection abundant.

It’s an incredible accomplishment, but abundance changes value. Polish once proved something because polish was expensive. Now it doesn’t.

So we are beginning to search for different signals. Not necessarily for things that are rougher or slower, or some romantic return to a world before technology. We’re searching for evidence of investment. Evidence that someone was there.

The organizations that understand this won’t reject optimization. They’ll become much more deliberate about where optimization belongs and where it quietly strips away something people value. Because in a world where almost anything can be made to look perfect, perfection is no longer proof that something mattered.

Increasingly, the proof will be what someone was willing to put into it.

And knowing when that investment is worth making will require something no machine can guarantee in advance.

Courage.