The screen glowed with a pale, surgical blue. It was three in the morning in a cramped apartment in Lyon, and Anne Alombert was staring at a digital manifestation of our collective future. On the monitor was the latest iteration of an artificial intelligence system designed to optimize public services. It promised speed. It promised efficiency. It promised to eliminate the messy, unpredictable friction of human interaction from the machinery of governance.
To the engineers in California or the technocrats in Paris, this was progress. To Alombert, a philosopher who spends her life examining how tools reshape human thought, it looked like a tragedy in slow motion.
We are currently living through a quiet, bloodless coup. It is not being fought with soldiers, but with code. Day by day, decision by decision, we are outsourcing the vital organs of our society—our schools, our hospitals, our legal systems—to automated platforms managed by a handful of private monopolies. We have been told that this is inevitable. We have been conditioned to believe that algorithmically driven governance is superior because it is objective.
That is a lie.
When we replace a human bureaucrat, a teacher, or a social worker with an AI interface, we are not just upgrading our technology. We are downgrading our humanity. We are substituting the deep, contextual understanding of a living person with a statistical prediction model. The consequences of this shift are not just administrative; they are existential.
The Fiction of the Perfect Algorithm
Let us invent a citizen to understand what is truly at stake. We will call him Pierre.
Pierre is sixty-two, recently widowed, and trying to navigate the labyrinth of the state pension system after a lifetime of irregular freelance work. In the old world, Pierre would walk into a physical office. He would sit across from a woman named Martine, who had worked for the department for twenty years. Martine would look at Pierre’s chaotic paperwork, see the panic in his eyes, and understand the unspoken reality: this man is terrified of slipping into poverty.
Martine might spend an extra twenty minutes digging through an old archive, finding a loophole, or simply offering a word of reassurance that makes the institutional coldness bearable. That interaction is public action in its purest form. It is an act of care mediated by the state.
Now, consider Pierre’s experience in the optimized, AI-driven alternative.
He logs onto a portal. He is greeted by a chatbot. The chatbot does not have eyes; it has parameters. It processes Pierre’s irregular income statements through an optimization model trained on millions of standardized data points. Pierre’s life does not fit the template. The system flags his application as an anomaly. There is no button to explain the nuance of his freelance contract from 1994. There is no human to hear his voice tremble. The system simply delivers a calculated, unappealable "No."
The engineers would argue the system worked flawlessly. It followed the rules. It saved the state money. It was perfectly efficient.
But it was completely blind to justice.
This is the central tension that Alombert highlights in her critique of our current technological trajectory. When artificial intelligence is used to manage public action, it reduces complex human lives into flat data. It assumes that because something cannot be quantified, it does not exist.
The Digital Erasure of Knowledge
The danger extends far beyond automated bureaucracy. It threatens the very transmission of human capability.
For centuries, human culture has evolved through a process that philosophers call exteriorization. We take our internal thoughts and memories and place them into external objects. We wrote on clay tablets, then in books, and eventually on digital hard drives. Every time we externalize a mental process, we free up cognitive space to think new thoughts.
But there is a catch. If we externalize the process entirely without understanding how it works, we lose the capacity to do it ourselves. We suffer from cognitive atrophy.
Think about the simple act of navigating a city. Before smartphones, you had to build a mental map. You looked at landmarks, you made mistakes, you understood the spatial relationship between neighborhoods. Today, you follow a blue dot on a screen. If the battery dies, you are effectively lost in your own hometown. You have outsourced your spatial intelligence to a private corporation.
Now, project that atrophy onto the scale of an entire civilization.
What happens when judges rely on AI to determine recidivism rates and sentencing guidelines? What happens when doctors rely on diagnostic algorithms to the point where they no longer trust their own clinical intuition? What happens when teachers use automated systems to grade essays, grading students on their ability to mimic the very algorithms doing the evaluation?
We are systematically destroying human expertise in the name of convenience. The irony is exquisite and terrifying: we are spending billions of dollars to build machines that simulate human thought, while simultaneously designing an environment that forces humans to think like machines.
The Architecture of Control
The current conversation around AI is dominated by a false binary. On one side are the doomsayers, warning of a rogue superintelligence that will wipe out humanity. On the other side are the techno-optimists, promising a utopia of limitless productivity.
Both narratives are distractions. They draw our attention away from the real, immediate threat: the colonization of the public sphere by private infrastructure.
Public services exist to serve the common good. Private technology companies exist to maximize shareholder value through data extraction and user retention. These two objectives are fundamentally incompatible. When a government decides to build its public infrastructure on top of proprietary AI models, it is ceding its sovereignty. It is turning citizens into users, and public policy into a monetization strategy.
This is not a hypothetical dystopia; it is happening right now. Look at the way public discourse is managed on social platforms, where algorithms designed to provoke outrage dictate what information rises to the top. Look at how gig-economy platforms use algorithmic management to control workers without providing the traditional protections of employment.
When the state adopts these same tools to manage public action, it adopts the logic behind them. It becomes an instrument of automated discipline rather than social cohesion. The citizen is no longer an active participant in a democracy, but a passive data subject to be managed, predicted, and optimized.
Reclaiming the Public Square
How do we fight an adversary that is invisible, everywhere, and disguised as progress?
The answer is not to become Luddites. It is not about smashing the servers or retreating into a pre-digital past. Technology is an intrinsic part of what makes us human. We cannot separate our development as a species from the tools we create.
The real task is to change the nature of the tools themselves. We must insist that public technology be built for public action, under public scrutiny, and guided by human values.
This requires a radical shift in how we design and deploy these systems. If an AI tool is to be used in a school, it must be designed by teachers and students, not by a venture-backed startup in Silicon Valley. If an algorithm is to assist in the distribution of social benefits, its code must be open, its assumptions must be transparent, and its decisions must be fully reviewable by a human being who has the authority to overrule it.
We need to treat digital infrastructure the same way we treat physical infrastructure. We would never allow a private, foreign corporation to build our roads, decide who is allowed to drive on them, and secretly change the traffic laws every Tuesday based on an optimization algorithm. Yet, we have allowed exactly that to happen to the digital highways that govern our thoughts, our economy, and our institutions.
The Human Premium
Ultimately, the defense of public action against the encroachment of total automation is a defense of the unpredictable.
Human life is messy. It is full of contradictions, sudden shifts, and flashes of unquantifiable genius. Our best institutions are those that leave room for that messiness—that allow for the exception to the rule, the moment of unexpected empathy, the radical reassessment of a broken law.
Algorithms cannot handle the exception. They operate on probabilities derived from the past. They are inherently conservative, trapping us in a loop of our historical data. They can never invent a new way of being, because they can only project the patterns of yesterday into the tomorrow.
Consider Martine, our hypothetical social worker, one last time.
She did not just process Pierre's paperwork. She looked at him. She recognized his dignity. In that fleeting, mundane moment in a drab government building, she affirmed his place in society. That act of recognition cannot be coded. It cannot be optimized. It cannot be scaled.
It is the valuable thing we possess. And if we surrender it to the machine, we will find that the society we have automated is one we no longer wish to inhabit.
