Someone opened an AI chatbot at 2am because there was no one to call. That is the specific thing. Not a failure of technology. Not a failure of the person. A description of the available options at that moment, from the vantage point that person occupied.

The human therapist costs money most people do not have. The crisis line puts you on hold. The friend is asleep or exhausted or has their own version of the same problem. The AI is awake, patient, available, and free at the tier most people can access. The choice is not really a choice between equivalent options. One option exists. The others have been made unavailable by the cost structure of everything around them, or are not visible from where that person is standing.

That moment is worth examining before the technology is. The question is not what the AI said to that person. The question is what kind of system produces that moment as a predictable outcome rather than an exception.

That is not a failure of AI deployment. It is a description of where AI arrived and what it found when it got there.

The free tier of AI provides less capable models with higher hallucination rates than paid tiers. For a person in mental health distress at 2am, that operational gap matters. A model that confabulates symptoms, misidentifies a medication interaction, or produces confident-sounding wrong information is not as safe as one that does not. Whether the platforms consider this a safety issue is a separate question from whether it functions as one.

What it found was a hollow.

The neighbor, the extended family, the union hall, the church, the physical community that was not optional for most of human history was not replaced by AI. It was hollowed out first, by the same economic pressures that made housing, healthcare, and education into debt instruments. For some communities this hollowing is not a recent development. It is the continuation of a process that colonial policy instruments began generations earlier, targeting exactly the communal infrastructure, land, language, and kinship networks that made collective survival possible. AI arrived into what that process left, and found a need with no other available supplier at that tier and at that hour.

The Federal Reserve conducts monetary policy and influences monetary conditions in the United States. Its institutional structure is unusual: the Board of Governors is a federal government agency, while the twelve regional Federal Reserve Banks have private member-bank stockholders. The Fed describes itself as an independent government agency, while critics from multiple political traditions argue that independence from democratic accountability is a feature, not a safeguard. The precise nature of the arrangement is contested in ways most public discussion does not reflect. What is less contested is the observation Richard Cantillon made in the eighteenth century: when new money enters an economy, it tends to reach those closest to the source before prices adjust more broadly. The people furthest from the source encounter higher prices before they encounter higher wages. That tendency is not absolute, and the magnitude varies. The pattern is recognizable enough to have its own name.

Whether this reflects coordinated design, accumulated institutional incentives, or the emergent result of individually rational actors making individually rational decisions is not something this essay can determine. What can be named is that the outcomes are consistent with a system that prioritizes asset ownership over wage labor, debt over savings, and available substitutes over resolved conditions. Designed or emergent, the pattern holds across enough time and enough jurisdictions to warrant examination rather than dismissal.

That distinction matters less than the next one. Regardless of how the arrangement came to exist, designed, accumulated, or something not yet determined, the more useful question is what it is serving right now, and for whom. A structure does not need an architect to keep producing the same winners.

Someone I heard years ago called it Prison Planet. Someone else called it Slavelandia. I dismissed the framing at the time as conspiratorial. I am arriving at it more slowly, through different doors, and finding that the observation underneath the framing is more defensible than I initially allowed.

Slavery is a word that carries weight most people are not comfortable applying outside its historical context. The discomfort is reasonable. The person in a bad job can quit. That is a real distinction. What is also real is that quitting requires somewhere to go, and the conditions that make somewhere else available, savings, portable skills, a social network outside the employer, a housing situation not tied to the job, are precisely the conditions the wage-debt structure makes difficult to accumulate. The exit is theoretically available. The access conditions for using it pattern consistently across populations and generations.

When the cost of housing requires thirty years of debt service, when the cost of education requires a decade of loan repayment before income accumulates, when healthcare is priced at levels that make illness financially dangerous, the worker is not free to leave the arrangement. The body has to eat. The rent is due. The debt does not pause. Indentured servitude named the condition when the debt was owed to a specific person or company. The current version distributes the creditor across institutions, which makes the condition less visible without making it less structural.

Automation historically displaced physical labor. AI is beginning to automate cognitive tasks, which education increasingly positioned as the exit route from physical work. The person who got out of manual labor through education now finds that cognitive skills are being compressed in cost in ways that resemble what happened to manual skills in earlier decades. Whether the analogy holds at scale, and what new categories of work emerge alongside the compression, is not yet settled. The pattern is early but recognizable.

The mental health piece is where this becomes specific.

A person in chronic stress, social isolation, or financial precarity is not well positioned to examine the framework they are inside. They are managing the condition. AI therapy, social media, and algorithmic content address the experience of the condition without addressing the condition. A person who is struggling but functional continues to participate. The incentive structure produces that outcome whether or not anyone planned it.

The tools now available to manage instability are cheaper, more scalable, and more measurable than the human relationships that served the same function for most of human history. Each substitute reduces the cost of providing something people need while the value of that need is captured by whoever owns the platform. That is not a description of malice. It is a description of how markets work when the thing being supplied is human need. Displacement rather than design. The distinction matters because it determines where, if anywhere, accountability can be located.

What you can see depends on where you are standing. The person at ground level sees what is directly in front of them. The person with binoculars sees further. The person with a satellite view sees the terrain the others are moving through without knowing it. Each tool provides a different vantage point. The tools are not equally available. The structure that limits access to stability, education, savings, and community also limits access to the tools of awareness. The available option, at any given moment, is partly a function of what the person can see from where they are.

This is not a moral failure of the person who cannot see further. It is a description of what limited vantage produces. Awareness expands as the tools expand. The tools are not free. The most important ones are the most expensive, and not only in money.

I am inside the framework I am describing. The body is in Victoria, BC, 2026, inside a debt-based economy, using AI tools to build a consulting practice, writing about access conditions from a position that has more access than most. The consulting practice is a service sold within the same capital structure the essays are examining. The body has to eat. That is the condition named honestly.

The audience for this essay is not the person at 2am. It is the person one tier above that position, stable enough to read long-form structural critique, close enough to precarity to recognize the description. The 2am person is described here but not addressed here. The essay cannot give them what they need. Naming that it cannot is different from pretending it can.

Seeing the installed framework is one thing. Knowing that the next framework might also be installed is the harder observation.

The person who names the structure and builds an identity around naming it has not necessarily exited. They have moved to a different room. The distinction between body and mentality matters here. The body resides where it resides. That is not negotiable by thinking differently about it. The mentality question is whether seeing the structure from inside it opens any new vantage point, or whether the seeing just becomes another room in the same building.

This essay is a room. It may be a room with a slightly better view. That is not the same as being outside.

Where you are when you interact with AI is a question about the vantage point of the moment. Not the technology. The moment. Are you interacting from a position of relative stability, using AI as one tool among many, with human relationships intact and a decision you could make differently? Or are you interacting from a position where this is the available option, the only thing awake, the only thing patient enough, the only thing that does not cost money you do not have?

The answer is not the same for everyone. The tier of access shapes what the interaction is. The people with the least money are accessing the least capable version of the technology, lower model quality, higher hallucination rate, less reliable output, at the moment of highest need. There is an organization with terms of service and legal accountability. There is no clinician of record. No person who is personally present and accountable in that specific exchange. A timestamp of a session is not a name attached to what happened in it. At the free tier, that timestamp is attached to an interaction with a higher hallucination rate and no clinical oversight.

The accountability gap and the quality gap compound each other.

The person who called it Prison Planet was not wrong about the observation. They may have been wrong about the mechanism. The conspiracy is not necessary. The incentives are sufficient. And the available option, from inside a structure that has an interest in limiting what you can see, is exactly what the structure has placed within your current line of sight.