Purpose in the Age of Intelligence

Originally published on Medium.

AI is a sharp knife. It can be used for harm, or it can be used for benefit. That is the nature of all tools. They may imply a purpose, but they are not purposeful on their own.

The question has never been whether AI is powerful. It already is. The question has never been whether AI will change the world. It already has. The real question, the only question that actually matters, is whether we will have enough purpose, discipline, and courage to use it without surrendering the very humanity that made it possible.

Three Principles Before Everything Else

Before discussing what’s going wrong, I want to establish the framework I use to evaluate what’s going right. These three principles are the lens through which every AI adoption decision should be measured.

People first. If AI adoption moves people’s lives from a better state to a worse one, that is not a business win. It is a signal that the entire vision needs to be rewritten.

Human-in-the-middle. Every AI-powered process must keep a human at the steering wheel, not as a manual override when something breaks, but as a constant and intentional presence within the system. Not because AI can’t steer. In many cases it can, and does it better. But because building self-sustaining thinking systems without human accountability is the single largest unaddressed risk in modern technology. A representative of humankind must always be in control.

Safety as a boundary, not a feature. Wherever AI is implemented, human safety is the highest priority. Not a compliance checkbox. Not something added after the product ships. Safety is the boundary within which AI is allowed to exist at all.

Who’s in the Room

When I look at the AI landscape today, I see three groups.

The first is fully bought in, convinced that AI solves everything, moves fast, and that anyone questioning the pace is simply in the way. The second is cautious, interested, approaching carefully, watching what happens to the people around them before committing. The third is quietly opposed, whether or not they say so publicly.

I am not in the first group, though I believe AI is one of the greatest inventions in human history. The invention itself is not the problem. The adoption is.

What troubles me is watching all three groups react to forces they did not choose, with no shared framework to evaluate decisions against. The principles above are my attempt at that framework. Let me show you why we need it.

The Misunderstanding at the Top

Klarna made headlines in 2024 when its CEO announced that AI had replaced the equivalent of 700 customer service workers. The company framed it as an efficiency story. What the story actually illustrated was a dangerous misunderstanding of what a business is.

A business cannot endlessly remove people from earning and still expect people to consume. When you automate your workforce out of existence, you are not just cutting costs. You are eroding the market that buys your product. The customers Klarna serves are workers somewhere. The workers it replaces are customers somewhere else. This is not a philosophical concern. It is an arithmetic one.

UPS announced cuts of 20,000 workers in 2025. Microsoft laid off over 15,000 people, largely engineers, the same quarter it reported 13% revenue growth and confirmed that AI now writes roughly 30% of its code. Block announced plans to eliminate 40% of its global workforce in early 2026. The pattern is consistent: strong earnings, record AI investment, and simultaneous displacement of the people who built those earnings.

The CEO who makes this calculation is thinking about doing the same thing with less. The correct question is how to do more with the people already in the room.

The Reskilling Illusion

The standard response from major AI providers to displacement is that laid-off workers should “reskill.” New opportunities will emerge. AI will create jobs we cannot yet name.

Look at that claim carefully.

Reskilling requires time. It requires learning materials, financial stability, and most importantly, guaranteed outcomes at the end of the path. None of those conditions exist at scale for the people most affected. The World Economic Forum’s 2025 Future of Jobs Report projects that 92 million jobs will be displaced by 2030 while 170 million new roles emerge. That sounds like a net gain until you ask: who gets those new roles? The answer, based on current data, is that 77% of emerging AI positions require advanced degrees. The parent who just lost a customer service job at 43 does not have a path through that door, and the math does not care.

“Reskill” is something you say when you want the conversation to end without taking responsibility for it.

But here is what real reskilling actually looks like, and it has nothing to do with chasing the credentials AI companies dangle. Learning new skills is a fundamental part of human evolution, with or without AI. That part is not optional and never has been. What changes is what we are learning for. The investment should go into better decision-making, better tooling, and most critically, the full, practiced capability to perform a manual override on whatever AI task is running today.

That last point is not theoretical. Twenty-plus years in the software industry teaches you one thing with absolute certainty: systems always go down. Every AI provider, no matter how well-resourced, will experience outages. If the humans inside an organization have been so thoroughly offloaded onto AI workflows that no one remembers how to do the work manually, a single outage does not create an inconvenience. It creates a complete halt. The dependency becomes the vulnerability, and the vulnerability becomes the crisis.

Real reskilling prepares people to work with AI at full capacity, and to work without it when necessary. Anything less is not a workforce strategy. It is a single point of failure waiting to be discovered.

What Good Looks Like

The counterevidence is real, and it deserves equal attention. Industries that have adopted AI as a complement rather than a replacement have seen a 10% productivity increase, 3.9% job growth, and 4.8% wage increases compared to less-exposed sectors, according to a 2024 labor economics study. This is not incidental. It is the direct result of treating AI as an expansion tool rather than a substitution strategy.

The distinction is simple in principle and difficult in execution. A business that uses AI to free its people from overnight shifts, unnecessary operational burdens, and repetitive work, so those same people can build more, think more clearly, and contribute at a higher level, is a business that has understood the technology. It is not thinking about fewer people doing the same things. It is thinking about the same people doing things that were previously impossible.

That is a different company. It is also a more durable one.

Where AI Adoption Actually Starts

It starts with culture. And culture means knowing not just how to use AI, but when and what for.

This is where the third group, the quietly skeptical, has something the first group lacks: instinct. The instinct that something is being offloaded that should not be. That using AI to write your own thoughts, make your own decisions, navigate your own relationships, and form your own judgments is not productivity. It is regression.

Using AI to do research, write an email, build an application, or eliminate grunt work is exactly what the tool is for. Letting AI assist your thinking is useful. Letting AI replace your thinking is dangerous. The line between the two is not always obvious, but it is always there.

If individuals offload their humanity to a machine, the market will follow. Providers will build systems for people who no longer want to think, feel, or connect. That is a market that consumes itself.

Ownership Is Not Optional

The piece most people skip is control.

There are significant communities and movements investing in local AI deployment, not only to protect privacy, but to keep AI at the level where communities can govern it themselves. This matters more than it sounds. When you depend entirely on a third-party provider for the tools that run your work, your decisions, and your access to information, you are not a customer. You are a dependency.

The relationship changes the moment the provider decides it can do what you do. That is when the tool you depended on becomes the system you cannot afford. That is when the partner that enabled you becomes the competitor that replaces you. That is when pricing, policy, and access are no longer in your hands.

Local AI matters. Community-level AI matters. Ownership matters. Not because every centralized provider is adversarial, many are not, but because dependence without control is not a strategy. It is a vulnerability.

What No System Can Threaten

There are those, governments, corporations, ideological movements, who believe that going all-in on AI gives them a decisive, permanent advantage. That enough compute, enough data, enough automation creates a kind of dominance that human beings simply cannot resist.

They are wrong. And the reason they are wrong is not technical. It is deeper than that.

There is something in human beings that operates entirely outside the domain of data. Call it instinct. The decision made with no evidence, no precedent, no model, that turns out to be exactly right, at exactly the right moment, in exactly the right place. The general who senses the enemy’s next move before the intelligence arrives. The founder who bets everything on an idea that every metric says will fail. The parent who walks into a room and knows, without a word spoken, that something is wrong with their child.

These are not failures of reasoning. They are a different kind of knowing, one that emerges from lived experience, accumulated feeling, and a relationship with the world that cannot be quantified. AI can approximate patterns. It cannot inhabit a life. It can process billions of data points. It cannot stand in a moment and feel the weight of it.

Any adversarial force that believes AI dominance is total dominance has already misunderstood what they are trying to dominate. You cannot automate sovereignty. You cannot train away the human capacity to sense what is true before the evidence exists to prove it. You cannot model intuition, because intuition is not a model. It is the residue of being alive.

This is not wishful thinking. It is the most grounded observation available: every civilization that has tried to reduce human beings to inputs in a system has eventually encountered the one thing no system can predict. The moment a human being decides, from somewhere beneath logic, that enough is enough. And acts on it. Correctly.

The real threat is not that adversarial forces deploy AI effectively. The real threat is that we voluntarily abandon the very faculties that make us ungovernable. That we outsource our instinct to algorithms and our judgment to dashboards, and in doing so, hand over the only advantage that was ever truly ours.

This is why the exploration of what it means to be human, our purpose, our meaning, our capacity for knowing without proof, has never been more urgent. Not as philosophy for its own sake. As survival. As resistance. As the clearest possible signal that the most powerful thing on earth is still a human being who has not stopped paying attention to themselves.

The Choice

The three groups I described at the beginning all face the same decision, just from different starting points.

The all-in group needs to ask whether speed is a substitute for purpose. The cautious group needs to stop waiting for someone else to define the framework and start building their own. The skeptical group needs to distinguish between the technology and the implementation, because rejecting AI entirely concedes the entire space to people with fewer scruples about how it gets used.

All three groups have a role. All three groups share a responsibility.

AI can help us build things that were previously impossible. It can give back time, reduce friction, amplify human capability, and create space for the kind of work that is genuinely irreplaceable: creativity, judgment, care, and the lived experience that no model can manufacture.

But if we use it to offload our thinking, we lose direction. If we use it to offload our judgment, we lose taste. If we use it to offload our relationships, we lose connection. And if we use it to offload our humanity, there is nothing left to optimize.

The technology is not the threat. The abdication is.

And the antidote to abdication is simple, if not easy: stay curious about what you are, keep exploring the instincts and capacities that exist beneath your reasoning, and never fully hand the wheel to any system, no matter how capable, that has never had to live with the consequences of being wrong.

That is not a limitation on what AI can do. It is the foundation on which everything it does becomes meaningful.