AI Did Not Create Corporate Opacity. It Only Made It Impossible to Ignore

I am grateful for the technology. I am far less charitable toward the companies that package it, explain it badly, hide behind it, or congratulate themselves for sounding more responsible than they are.

For a while, the public conversation around AI was framed as if the central questions were mainly technical. Can the model reason? Can it summarize? Can it write code? Can it hallucinate less? Can it do in seconds what used to take hours?

Those are real questions. Some of them are genuinely fascinating. Some matter a great deal.

But they are not the whole story, and they may not even be the most revealing part of it.

The technology is not the disappointment. The companies are.

What AI has exposed, with uncomfortable clarity, is an old corporate habit: promising one experience at the surface while operating by another logic underneath. The more complex and consequential a product becomes, the more that gap matters. A product can still be useful. It can still be impressive. It can still solve real problems. But once the explanation stops matching the mechanism, trust begins to decay.

That is not just a communications problem. It is a structural one.

The Problem Is Not Limited to “AI Companies”

It would be neat, and therefore dishonest, to frame this as a problem unique to AI companies.

It is not.

The companies building AI products do it. The companies quietly inserting AI into existing products do it. The companies loudly insisting that they do not use AI often do a version of it too.

Some wrap themselves in innovation theatre. Others wrap themselves in moral theatre. Both can be forms of evasion.

One company says, “Look at our extraordinary new intelligence.” Another says, “Unlike them, we remain human-first, thoughtful, restrained, untouched by the machine.” Neither posture guarantees honesty. Spectacle is not honesty. Restraint is not honesty. Branding rarely is.

The real divide is simpler than that. It is the divide between companies that explain the real exchange and companies that still believe opacity is sophistication.

AI has simply made that divide harder to ignore.

Because once a product becomes more dynamic, more language-shaped, and more entangled with user judgment, the old habits begin to look worse. Companies promise more than they explain. They describe benefits more clearly than limitations. They talk about safety in abstractions. They bury material caveats in secondary documentation. They shift the burden of understanding onto the user. And when criticism arrives, they retreat into scale, complexity, innovation, or inevitability.

None of this was invented by AI. AI just shines a floodlight on it.

Businesses Tell Stories Too

This matters to us at Kanonyq because we work on fiction, but the principle is larger than fiction.

Every business tells a story. Not metaphorically. Structurally.

A company tells the customer what the product does, what it does not do, what kind of outcome to expect, what kind of uncertainty remains, where the limits are, what the risks are, how much control the user retains, and what happens when the system fails. That is narrative structure whether the company likes the term or not.

A business makes promises. It creates expectations. It establishes cause and effect. It teaches the customer how to interpret the experience. It sets up payoffs. It trains trust.

And just like a novel, a company begins to lose trust when the underlying structure stops honouring the promises it has taught the audience to believe.

That is why so much corporate language feels wrong before a customer can always explain why. The surface may be polished. The interface may be elegant. The brand voice may be careful, warm, modern, and faintly anaesthetised. The help centre may look pristine. The product pages may glow with competence.

And yet something underneath feels off.

Why?

Because the story being told at the surface is not the same story being lived at the structural level.

The promise says clarity. The system runs on ambiguity.

The promise says control. The practical experience requires the user to discover hidden conditions, implied trade-offs, and operational limits after the fact.

The promise says transparency. The mechanism says: we disclosed enough, somewhere, for our own purposes.

That gap is where trust starts to die.

The Performance of Transparency

One of the most exhausting features of contemporary corporate language is that companies have learned to perform transparency without practicing it.

They tell you they value openness. They tell you they are committed to trust. They tell you they care deeply about safety, privacy, integrity, responsibility, empowerment, or whatever other polished noun the season requires.

Then they make the customer do archaeology.

You are given a homepage promise, a product page promise, a help-centre explanation, a privacy policy, terms of service, a few UI hints, perhaps a settings panel, perhaps a blog post, perhaps a buried FAQ, perhaps a toggle hidden behind an icon apparently designed by a committee of apologetic ghosts.

You are informed, technically speaking.

But you are not clearly told the thing you actually need to know.

That distinction matters.

A customer does not need corporate literature. A customer needs orientation.

What does this product actually do in ordinary use? What are its failure modes? What costs am I absorbing that are not obvious at first glance? How much of the burden of correction falls on me? What happens if I rely on this too much? What behaviour should I expect from the system in practice, not in polished abstraction?

That is the standard.

Not whether a company can point to a paragraph three clicks deep and say, “Well, technically, we mentioned that.”

The difference between honest explanation and defensive disclosure is not subtle. It is practical. It is ethical. And eventually it becomes structural, because it determines whether the user is operating inside a relationship of trust or inside a maze of interpretive labour.

That is not transparency. It is fog with legal formatting.

AI Raises the Cost of Evasion

AI raises the stakes because it changes the relationship between product and user.

A hammer does not pretend to understand you. A spreadsheet does not present itself as a collaborator. A conventional search engine may frustrate you, but it does not usually simulate judgment.

AI products are different. They operate in language. They imitate assistance. They enter cognitive work. They shape how users think, write, decide, revise, and trust.

That means the cost of opacity rises.

If a product can influence reasoning rather than merely execution, then clarity about its behaviour is no longer decorative. It is part of the product. It belongs on the critical path, not in the footnotes.

And yet many companies still treat it as a secondary layer. Something to soften. Something to patch later. Something to distribute across policy pages and settings menus instead of saying plainly what the user needs to know where decisions are actually being made.

Some companies organise everything around the spectacle of capability. Others organise everything around the reassurance of restraint. Both can end up doing the same thing: telling the user a cleaner story than the system deserves.

That is not a technical bug. It is a business choice.

The Companies That “Do Not Use AI” Are Not Automatically Better

There is another increasingly tiresome habit worth naming.

Some companies now present themselves as morally superior because they do not use AI, or because they use it quietly, or because they define themselves against the vulgar excesses of the current market.

Fine. Some of them may indeed be better. Some probably are.

But refusal to mention AI is not the same as corporate honesty. Distance from AI hype is not the same as respect for the customer. A business can avoid AI entirely and still communicate like a coward.

It can still hide limitations. It can still inflate promises. It can still bury risk. It can still make the customer carry interpretive labour that should have been handled by the company itself.

The medium changes. The habit does not.

That is why I am not interested in sentimental divisions between “good old-fashioned companies” and “bad AI companies.” Too much of that distinction is branding camouflage. The real question is simpler: does the company explain the product in a way that allows the customer to understand the real exchange?

If not, the problem is not merely technological. It is structural and ethical. It is a failure of explanation at the exact point where explanation becomes part of the product itself.

Where Trust Actually Breaks

At Kanonyq, we spend our time identifying where novels lose reader trust.

A story can have beautiful prose and still fail. It can have a strong premise and still fail. It can have compelling scenes and still fail.

What usually gives way is not ornament but structure.

Cause and effect weaken. Promises are introduced and not paid off. Characters stop earning the story they inhabit. Pressure dissipates. The system quietly begins to cheat.

The same principle applies outside fiction.

A company loses the customer when its explanations stop matching its operations.

When the promise and the mechanism diverge, trust does not usually collapse in one dramatic instant. It decays. The user feels the strain before they can always name it. They start compensating. They become cautious. They begin double-checking what should have been clear. They keep one hand on the door.

That is the damage.

And the most damaging companies are not always the loudest or the most ridiculous. Sometimes they are the most polished. The ones that know exactly how to sound clear while remaining materially evasive. The ones that understand how to soothe while withholding orientation. The ones that have mastered the appearance of responsibility without accepting the burden of plain explanation.

Those are the structurally broken stories.

Those are the dangerous ones.

What I Still Respect

I still respect the technology. That is not the disappointing part.

I respect the engineers, researchers, and builders who make astonishing tools possible. I respect the fact that many of these systems really are useful. They can help people think, compress tedious work, speed up analysis, explore alternatives, test language, accelerate iteration, and widen access to certain kinds of capability.

To deny that would be performative nonsense.

But precisely because the technology is powerful, the corporate behaviour around it becomes less forgivable, not more.

The more useful a tool becomes, the less acceptable it is to explain it badly.

I do not respect the corporate habit of treating explanation as a liability. I do not respect the belief that the customer should have to reverse-engineer the product’s practical truth from scattered hints. I do not respect the increasingly common performance in which businesses try to appear responsible by sounding careful while still preserving every advantage of ambiguity.

That is not maturity. It is just better-dressed opacity.

The Standard Should Be Higher

If your product affects how people think, write, decide, revise, trust, or work, then you owe them more than disclaimers and posture.

You owe them orientation.

You owe them the clearest account of the real exchange: the benefits, the risks, the limitations, the asymmetries, the practical costs, and the places where the system may fail them in ways they will otherwise discover too late.

Not because transparency is fashionable. Because without that clarity, trust becomes counterfeit.

And once trust becomes counterfeit, the relationship is already failing, no matter how good the interface looks.

That is true in fiction. It is true in technology. It is true in business.

AI did not create this problem. It only raised the cost of it.

What has been exposed is not merely a new class of tools, but an old corporate instinct: tell the customer just enough to keep the machine moving, and call the rest transparency.

I am grateful for the technology. I am not inclined to flatter the companies.

When the promise and the mechanism diverge, trust begins to fail. That is true in fiction. It is true in business. AI did not invent that fracture. It only made it easier to see.