The Document Nobody Wanted to Write or Read

A recent article about the workplace “AI doom loop” describes a scene that is already easy to imagine. Someone receives a document that appears to have been generated by artificial intelligence. The recipient reads it, senses that it contains polished language but little that is distinctive, and asks another AI system to summarize it. After one machine has produced the document and another has compressed it, almost nothing remains that might lead to a new thought. The words have moved efficiently from sender to recipient, but neither person has formed a meaningful relationship with them.

The warning deserves attention. Offices can generate enormous quantities of language without generating an equivalent amount of understanding. Reports, presentations, emails, proposals, and summaries may circulate because an organization expects visible evidence of activity. AI makes it possible to produce these materials at greater speed and scale. It can also help recipients avoid reading them. One person delegates the responsibility to articulate an idea, while another delegates the responsibility to interpret it. What appears to be a triumph of productivity may become a system for manufacturing and processing documents that nobody genuinely wanted to write or read.

Yet the failure in this scene does not arise from AI alone. Long before generative systems became widely available, institutions produced documents filled with fashionable terminology, predictable recommendations, and conclusions designed to offend no one. Employees copied passages from earlier reports, adjusted presentation templates, and expanded simple observations into pages of professional language. Recipients skimmed executive summaries, searched for action items, and ignored most of the supporting text. AI accelerates an existing tendency, but it did not invent language without engagement.

The deeper problem is the withdrawal of human attention. The writer uses AI to avoid developing a position, and the reader uses AI to avoid encountering one. Between them sits a document whose polished surface conceals an intellectual absence. Calling this process a doom loop is appropriate because every stage reinforces the next. Generic writing encourages superficial reading, superficial reading reduces the demand for meaningful writing, and the resulting decline in attention makes further automation appear necessary.

That loop is real, but it is not the only relationship available between human beings and AI. The same technology can be used to remain with a question longer, examine assumptions more carefully, and explore connections that might otherwise be abandoned. A machine that helps someone escape thought can also help another person continue thinking. The difference lies not only in the capability of the system, but in the practice formed around it.

The Four Stories We Tell About AI

Public discussion tends to organize artificial intelligence through several recurring stories. The first is the story of the shortcut. A person enters a prompt and receives a finished report, slide deck, image, or campaign within minutes. Demonstrations emphasize the contrast between the old laborious method and the new effortless result. Their promise is direct: what once required hours can now be completed almost immediately.

This story captures a genuine advantage. Many tasks contain mechanical work that deserves to be reduced. Reformatting a table, correcting routine language, reorganizing notes, or preparing a provisional structure need not consume the best hours of a person’s day. Trouble begins when reduced effort becomes the entire definition of progress. The most desirable tool then appears to be one that allows us to avoid the greatest amount of involvement.

The second story concerns substitution. Here AI does not merely assist a worker but becomes capable of performing the worker’s role. Predictions focus on which occupations will disappear, how quickly organizations will reduce headcount, and whether human labor will retain economic value. Again, there are substantial reasons for concern. Automation has always changed the distribution of work, and contemporary AI can affect professional tasks once believed to require secure forms of expertise. Workers cannot be expected to treat that possibility as an abstract intellectual exercise.

The third story is apocalyptic. An increasingly capable intelligence exceeds human control, pursues goals that conflict with human survival, or gives dangerous power to governments, corporations, and malicious actors. Some risks deserve serious investigation, especially in cybersecurity, military applications, surveillance, political manipulation, and critical infrastructure. The problem arises when uncertainty is converted into dramatic inevitability. AI becomes a single force moving toward a predetermined end, while human institutions and choices fade from view.

These three stories share an image of the human being as increasingly absent. In the shortcut story, people escape effort. In the substitution story, they become unnecessary. In the apocalyptic story, they lose control of history itself. Each perspective identifies a real possibility, but together they leave little room for a different experience already taking place in ordinary life.

A fourth story begins with intellectual partnership. AI can help a person ask more precise questions, find relevant sources, compare interpretations, test an argument, and express an intuition that has not yet acquired a stable form. Its value is not exhausted by completing a task more quickly. It can expand the field within which the person thinks.

This partnership does not turn AI into an unquestionable authority. Its responses can be mistaken, incomplete, excessively agreeable, or seductively fluent. The user must examine sources, challenge conclusions, and retain responsibility for judgment. These demands do not make the partnership unsuccessful. They are part of what makes it intellectually productive. A useful interlocutor is not valuable because it eliminates the need to think, but because the exchange calls more thinking into existence.

From Saving Time to Expanding Thought

Productivity is commonly measured through a relationship between time and output. If a report once required eight hours and can now be produced in one, productivity has increased. If a writer once published one essay each month and can now publish several, the improvement appears equally obvious. Such measurements are useful, but they describe only the visible surface of work.

Time saved does not tell us what happened to attention. An employee may use the recovered hours to investigate the problem behind the report, or management may demand seven additional reports. A writer may explore more ambitious questions, or fill a publishing schedule with material that adds little to public understanding. The same efficiency can create intellectual freedom or intensify the machinery of production.

A richer idea of productivity asks what becomes possible within the time available. Research can extend beyond the first persuasive source. An initial opinion can encounter objections before it hardens into a public declaration. Historical background can reveal that an apparently new problem has older forms. A comparison that first seems attractive can be tested for limits. Language can be revised until it reflects what the writer actually believes rather than what merely sounds impressive.

This is exploratory productivity. Its achievement cannot always be counted in documents or minutes. It appears in the distance traveled between the first intuition and the final understanding. A person may begin with a simple reaction, follow it into several disciplines, discover that part of the reaction was mistaken, and return with a more responsible position. The process has been productive even if the number of published pages remains unchanged.

AI can support this movement because it lowers the cost of asking the next question. One answer can be challenged immediately. A forgotten name or historical connection can be investigated while the line of thought is still active. Several possible structures can be compared before the writer commits to one. A misunderstanding can be identified before it becomes the foundation of an argument. The resulting speed matters because curiosity often loses energy when every small investigation requires a long interruption.

The benefit is not automatic. Fluency can create the impression that an inquiry has been completed when it has barely begun. A plausible explanation may discourage verification, and a well-organized draft may make a weak idea feel mature. Intellectual productivity therefore requires a deliberate reversal of the shortcut mentality. Instead of asking AI to finish the work so that thought can stop, the user asks what else must be considered before the work deserves to be finished.

Quantity may increase as a consequence. A person who can move more readily from observation to research and from research to articulation may write more often. That increase does not necessarily reveal an obsession with output. It can indicate that the practice has become sustainable and that each completed inquiry supplies energy for another. Productivity, in this sense, is not the multiplication of content for its own sake. It is the continuity of an intellectual life.

The Long-Form Essay as a School of Attention

The contemporary media environment rarely rewards this continuity. Much of it is designed around small units that can be consumed quickly and replaced immediately. Short videos begin with a hook, accelerate toward an easily recognized point, and disappear as the next item enters the screen. Articles are divided into fragments for readers expected to scan rather than follow an argument. Lists promise useful knowledge without requiring prolonged attention to relationships among ideas.

Concise forms are not inherently shallow. A poem, aphorism, brief reflection, or carefully made video can remain in the mind for years. Compression can require severe intellectual discipline, and a short work may open questions far larger than itself. The difficulty lies in an economy of attention that values content primarily for its ability to produce an immediate response. Under those conditions, brevity is less an artistic choice than a strategy for preventing the audience from leaving.

The long-form essay creates another tempo. It asks the writer to remain with an idea after the initial excitement has faded. A reaction that seems convincing in a paragraph may become unstable across several pages. Examples must do more than decorate a claim. Connections require explanation. Counterarguments cannot be acknowledged through a ceremonial sentence and then forgotten. Length does not guarantee depth, but it creates space in which superficiality becomes harder to conceal from the writer.

Paragraphs are central to this discipline. A paragraph is not merely a visual break inserted to make a page easier to read. It is a unit of thought. It introduces a movement, develops its internal relationships, and reaches a point from which the next movement can begin. When a writer treats paragraphs this way, composition becomes a form of reasoning. A sentence that cannot find its place may reveal an incomplete connection. A paragraph that keeps expanding may contain several thoughts that have not yet been distinguished.

Sections perform a related function at a larger scale. They allow an essay to shift perspective without becoming a collection of unrelated observations. A personal experience can lead toward history, history toward a contemporary problem, and that problem toward a moral or philosophical question. Each change requires enough continuity for the reader to understand why the journey is being made.

The writer learns through this construction. What begins as a feeling must become an intelligible claim. What appears to be a private concern may reveal social or historical dimensions. An opinion may survive the investigation, but its reasons often change. At other times, the act of writing exposes a contradiction that conversation alone had not made visible. The finished essay records not only what the writer thinks, but the path through which the thought acquired greater clarity.

Readers are invited into the same movement. They do not merely receive a conclusion but accompany its development. A long essay gives them time to disagree, reconsider, connect the argument with their own experience, and recognize where their attention resists. Reading becomes participation in another person’s effort to understand.

AI fits this form better than the culture of instant output often suggests. Used well, it can support the research, questioning, and revision required to sustain a long argument. The long-form essay, in return, can discipline the use of AI. It refuses to treat the first fluent answer as sufficient and asks whether the relationships among ideas can endure extended examination.

A Chosen Form, Not an Assembly Line

My essays currently tend to follow a recognizable structure. They usually extend beyond 2,000 words and are organized into around seven sections. Each section marks a genuine movement in the inquiry, while each paragraph is expected to carry a complete development of thought. This regularity did not begin as a theory about how every essay ought to be written. It emerged through trial and error as a form that helps me learn through writing.

The consistency could create suspicion in an age when people search for signs of automated production. Readers may associate regular headings, balanced paragraphs, and sustained grammatical control with AI. Yet uniformity of structure does not establish uniformity of thought. Established forms have always helped writers concentrate their attention. Sonnets use a fixed number of lines. Academic articles follow recognizable conventions. Homilies, columns, and musical compositions develop within inherited patterns. Constraint can focus invention rather than prevent it.

A form becomes an assembly line only when it predetermines what every subject must become. If each section exists merely because a space must be filled, the writing begins to serve the template. If the sections create room for observation, context, objection, revision, and return, the structure serves the inquiry. The difference becomes visible in whether the writer discovers anything that was not already contained in the opening opinion.

The choice of around seven sections is not a mystical prescription, nor is the threshold of 2,000 words a universal measure of seriousness. Both represent my present understanding of the space a thought often needs. Another writer may work through a different architecture, and my own practice may change. A form remains alive when it can be revised in response to experience. It becomes oppressive when preserving the form matters more than following the thought.

For now, this structure gives ordinary observations a fair hearing. A childhood memory, a design preference, a sentence heard at Mass, a reaction to workplace culture, or a small discomfort with a familiar habit can become the beginning of a wider investigation. The subject does not need to prove its importance before writing begins. Significance may emerge because attention has been sustained long enough to reveal it.

This approach resists the demand that every idea immediately justify itself through popularity. A topic does not have to promise a large audience, fit a current controversy, or compress neatly into a slogan. Writing can begin because something remains unresolved in the writer’s mind. The essay then becomes the means of discovering why it mattered.

Authorship After the Myth of Solitary Creation

The growing presence of AI has encouraged some writers to emphasize that their work was created entirely without it. The statement may express a legitimate commitment to craft. Writing unaided can strengthen memory, patience, concentration, and sensitivity to language. Anyone who chooses that path deserves respect, especially when the choice is made thoughtfully rather than used as a claim of moral superiority.

The absence of AI, however, does not guarantee quality. A completely human-written essay can be derivative, careless, poorly researched, or governed by unexamined prejudice. Human beings have always been capable of producing slop. Machines have made its production easier, but they did not create the underlying temptation to speak without sufficient thought.

AI-assisted writing must face the same standard from another direction. A fluent draft may contain little of the person who publishes it. Sources may be attached without being understood. Apparent balance may come from predictable qualifications rather than a genuine encounter with opposing positions. When the writer cannot explain or defend the resulting argument, polished language becomes a substitute for authorship.

Hybrid work can also achieve a high level of intellectual and literary quality. A writer brings memories, questions, moral commitments, professional knowledge, and a sense that something deserves attention. AI can help test the connections, locate evidence, identify misunderstandings, and propose language. The writer accepts, rejects, reshapes, and remains responsible for the result. Such a process cannot be understood adequately through the binary categories of human or machine.

Authorship has never meant creation without influence or assistance. Writers think with books, teachers, editors, conversations, dictionaries, archives, and traditions of form. Even solitary writing occurs inside a language received from others. AI changes the speed and intimacy of assistance, which creates new responsibilities, but it does not erase the long history of thought developing through encounter.

Disclosure remains important where rules, professional standards, or public trust require it. Academic assignments, journalism, legal work, and organizational documents may depend on clear accounts of how material was produced and verified. Personal reflective writing occupies another context. Constantly announcing every tool can draw attention away from the ideas and create a misleading picture in which authorship is measured by technical isolation.

A more durable standard is responsibility. Does the writer understand the claims? Have sources been examined rather than used as decoration? Has the writer retained the freedom to disagree with the tool? Can errors be acknowledged and corrected? Is the published essay an honest expression of a considered position? A name beneath an essay signifies accountability for these decisions, not the absence of every form of assistance.

Beyond the Doom Loop

The doom loop begins when everyone wants the appearance of communication without the labor of attention. AI generates what the sender has not fully considered and summarizes what the recipient does not wish to read. Each cycle makes the next document less meaningful and the next act of reading less necessary. Productivity rises according to the measurements of the system while understanding declines.

Another loop is possible. An experience gives rise to a question. The question leads to conversation and research. New information complicates the initial reaction. Writing gives the developing thought a structure, while revision exposes gaps that require further inquiry. The completed essay changes how the writer encounters the next experience, which becomes the beginning of another question.

AI can participate in this cycle without becoming its author or destination. It can help preserve the momentum of curiosity, especially when a question crosses fields or requires knowledge beyond immediate recall. It can also introduce friction by challenging a claim, presenting a neglected perspective, or showing that an attractive comparison does not survive closer examination. Assistance does not always mean making the path easier. Sometimes it means revealing that the path is longer than expected.

This constructive account does not require technological optimism without limits. AI can mislead, intensify inequality, disrupt work, concentrate power, and strengthen systems of surveillance or manipulation. These risks deserve evidence, regulation, institutional responsibility, and continuing public scrutiny. Apocalyptic certainty is not required to take danger seriously. Fear becomes less useful when it treats human extinction as the only future worth discussing and overlooks the choices being made in workplaces, schools, homes, and creative practices now.

The brighter possibility is already modestly visible. People can use AI to study subjects they once considered inaccessible, communicate across languages, test ideas without waiting for formal access to an expert, and maintain intellectual projects that ordinary pressures might have interrupted. None of this guarantees wisdom. It expands the opportunity to pursue it.

Long-form writing can become one of the disciplines through which that opportunity is realized. It slows the movement from prompt to publication. It gives evidence, uncertainty, personal experience, and competing interpretations enough room to coexist. It asks both writer and reader to resist the assumption that the fastest answer is the most valuable one.

Writers do not need to reject new tools to protect thought. They need practices that keep thought present while the tools become faster. A chosen form, sustained attention, responsible use of sources, and willingness to revise can turn increased capability into intellectual growth. The result may include more essays, but their number is not the final measure. The better measure of productivity is not how quickly technology allows us to stop thinking. It is how much farther technology enables us to continue.

Photo by Annie Spratt on Unsplash