Category Archives: science, technology and society

what it would take for the public to trust science

On July 19, the President posted on the social media platform that he owns:

The National Academy of Sciences has historically published analytical Scientific Manuals expressly for our Federal Judiciary. Of course, the Academy has been run by Radical Left Dumocrats who, it turns out, published fraudulent, biased, and misleading Manuals on Climate Change. These bogus Manuals were used by Judges to decide massive “Climate Change” Cases, and have created huge losses across our Country. These Manuals have been totally DISCREDITED. Our Nation’s Federal Judges deserve Facts and Science, not Political Fraud and False Science on Climate. With this TRUTH, I hereby order Federal Suspension and Debarment Officials to review this conduct. Our Taxpayers should not be funding Climate Fraud, and Judges should never have relied upon it. Thank you for your attention to this matter! President DONALD J. TRUMP

When I heard about this post, I assumed that I would trust the National Academies manual that Trump was denouncing. This was an interesting reaction because I had not seen the document.

I then looked at it (and you can too). It struck me as cogent, informative, and cautious. But I do not independently know whether its claims are true.

Nor do its authors. For the most part, the document refers to large, organized groups, such as the Intergovernmental Panel on Climate Change and the National Research Council. But consider a scientist whose work is very close to the natural phenomena: collecting and measuring atmospheric samples. This person knows what they are doing but not what everyone else’s samples show, let alone how to build the instruments that they all use. And current CO₂ concentrations are only some of the variables that we must measure to understand the climate. The vast majority of what any individual scientist knows comes from reports of what others have done.

As the report that Trump denounced says,

The Earth’s climate system is enormous and complex, consisting of many nested and interlinked subsystems, including the atmosphere, hydrosphere, cryosphere, lithosphere, and biosphere. …. Owing to the breadth and complexity of the climate system, climate research spans many different scientific disciplines. For example, research on the physical science of climate and climate change (i.e., understanding of the physical properties of the climate system and how it is changing) encompasses: (1) mathematics and statistics; (2) basic sciences, e.g., physics and chemistry; (3) earth sciences, e.g., atmospheric sciences, climatology, physical geography, oceanography, meteorology, hydrology, biogeochemistry, and cryospheric sciences; and (4) computer sciences and data analysis. Most climate research involves collaboration across these different fields (National Academies 2025)

The chemists must trust the oceanographers, who must trust the computer scientists. In fact, each chemist must trust the other chemists.

Perhaps all of science depends on trust, although that is less obvious in classic cases, like Galileo dropping objects from the Tower of Pisa, than in climate science. The study of the earth’s climate is intensely collaborative, involving many more people than could ever know each other. It depends on the Integrated Assessment Models, which have developed for more than fifty years (Van Beek, Hajer, Pelzer, van Vuuren & Cassen 2020). Anyone who contributes to these models must presume that the previous contributions have been generally reliable, even when new data changes some of the current findings.

Why do I trust those models? Why do I trust the National Academies summaries of them, or Karen Zraick’s New York Times article that drew my attention to Trump’s “TRUTH”?

On one hand, I have no doubt that the cautiously presented conclusions of the National Academies are valid. On the other hand, my confidence has a sociological explanation. Institutions like universities, scientific disciplines, and the Times have treated me well all my life. I have benefited from operating with an overall model that incorporates confidence in what these institutions say plus direct observations that concur with their findings. I have observed the climate get hotter in New England, but I only believe that this is a global trend and a function of humans’ burning carbon because I believe in institutions, like The New York Times, that report on institutions like the National Academies.

Few people are treated as well or have as much experience with these institutions as I do; I am a college professor in the United States. It is unreasonable to ask people to trust institutions that are unaccountable to them and distant from them. I think that scientific institutions can adjust their behavior to be somewhat less distant from the public, but they cannot bear that responsibility alone.

The classic solution has been intermediary organizations that obtain people’s trust and then independently present information. An important example was the 20th century city or town newspaper. In 1972, the General Social Survey (GSS) found that 69 percent of Americans read a newspaper daily.

Some 20th century US newspapers had specific social or ideological slants, but the market was dominated by local monopoly publications. Like the broadcast television networks of the same period, newspapers would have presented the American Academies’ position on the climate as scientific truth. In those days, Trump’s response would have been much more marginal.

In 1998 and 2008, the GSS asked people whether we trust too much in science. In both years, those who never read a newspaper were much more likely to distrust science than those who did read one. When I made a model that controled for age, race, gender, and years of education, reading the newspaper still predicted less distrust in science, net of the other factors (p <.001).

This doesn’t mean that newspapers caused people to trust science. It could be the other way around–people who trusted science subscribed to newspapers. But I think that newspapers devoted resources and skill to building trust and then guided people’s reactions to other institutions, such as science.

The recipe that sustained metropolitan daily newspapers during the 1900s included hard news plus human-interest stories, comics, sports, gossip, and other features aimed at various members of each household, all funded by a combination of revenue from subscribers and advertisers. Once the internet captured advertising revenue, this model disintegrated. By 2014, just one-quarter of respondents told the GSS that they read a newspaper daily, and the GSS subsequently retired the question, as it ceased to be a useful measure of news consumption.

All intermediary institutions have “business models” of some kind. A 20th century newspaper relied on subscribers and advertisers. An American high school typically gets most of its resources from local property taxes in return for training and minding the community’s teenagers and providing sports and entertainment. Among other things, a school teaches scientific findings and respect for science. Similarly, a public university in the United States offers job training and credentials, health services, quasi-professional sports, patented inventions, and high culture in return for tuition, donations, state subsidies, and fees. Meanwhile, the big social media platforms use algorithms to attract attention, which they sell to advertisers.

Each business model determines the institution’s results, including its impact on trust or distrust for science. Metropolitan daily newspapers tended to boost trust, while today’s social media degrade it.

Trust in science is not self-evidently good; in fact, we can trust too much. And the fact that most newspapers were local monopolies was problematic. They generated trust, but arguably at the expense of freedom. The question is whether we can invent business models for the 21st century that yield genuinely civic outcomes.

Sources: National Academies of Sciences, Engineering, and Medicine. 2025. Reference Manual on Scientific Evidence: Fourth Edition. Washington, DC: The National Academies Press; Van Beek, L., Hajer, M., Pelzer, P., van Vuuren, D., & Cassen, C. (2020). Anticipating futures through models: the rise of Integrated Assessment Modelling in the climate science-policy interface since 1970. Global Environmental Change, 65, 102191.

paradigms of attention

Leave a reply

Attention as selecting information to accomplish a task

Imagine that you run a laboratory in which you give people or animals stimuli (for instance, showing them shapes on a screen or making sounds) and you record their responses.

Sometimes, your subjects miss things in interesting ways. For example, in the Invisible Gorilla experiment, viewers are told to count how many times players pass a basketball, and most fail to notice a person in a gorilla suit who walks across the court (Chabris and Simon 2011).

This general approach dates back at least as far as Wilhelm Wundt’s psychology lab in Leipzig in the 1880s. Since World War II, psychologists have often posed an analogy between their subjects and computers–specifically, the large class of computers called “von Neumann Machines” (van Neumann 1945). These familiar devices have components that collect data (keyboards, cables, cameras), components that apply stored instructions to process data, components that store data, and components (such as printers and screens) that output the computed data.

Von Neuman sketched the design of a “very high speed automatic digital computing system” and then directed the team that built a prototype computer at the Institute for Advanced Study from 1948-51. According to his sketch, this machine’s computing function, its command function, and its memory function would “correspond to the associative neurons in the human nervous system.” The equivalents of our sensory and motor neurons would be “the input and the output organs of the device” (2.6).

If indeed a person or an animal is like a von Neumann machine, and it fails to notice something, the likely causes would seem to be a limitation of sensory function, a scarcity of computing power, or a failure of memory-storage. For example, if we miss the guy in the gorilla suit–as I did when I was first shown the video–it’s either because something was wrong with my eyes, I was focusing all my processing power on counting the basketball passes, or (conceivably) I saw the guy dressed as a gorilla but I failed to remember him. Most explanations emphasize limitations in our processing power, or “information-bottlenecks.” Then “attention” is whatever process we use to direct our finite processing capacity.

Attention as care for another person

Now imagine that you are observing two friends in a conversation. One tells the other a sad story, and you think that she seeks an expression of sympathy or compassion from the other. But instead of displaying any emotion in response to the story, the other begins to talk about himself.

Here it is unlikely that we will think of a computer or an information-bottleneck. The problem is not that Harry is receiving too much data to be able to process Sally’s story. The problem is motivational. Harry does not care enough about Sally, possibly because he cares too much about himself.

Von Neumann machines have no motives and do not care. Their designers do. Engineers can make machines that scan patients to assist with medical care or that identify enemy soldiers to destroy them. Such differences are crucial, but they are not caused by the machines. They are encoded in the machines’ design, which could be different and which can be changed. Further upstream is an institution, such as a hospital or an army, that paid for (or ordered) the machine to be built. Behind the institution are the people who designed, supported, and led it.

Thus, when we think about this second case, we are likely to ask what Harry and Sally want and possibly what the culture and society have taught or encouraged each of them to do–not about their cognitive limitations.

In fact, people can overcome their inability to grasp or process another’s story by devoting more time to it. In lab experiments, reaction-time is often treated as evidence of attention, but in relationships, the time that we devote to attending anything or anyone is a variable under our control. Indeed, some people choose to devote their whole lives to attending to one thing.

Attention as respect for something beyond me

Another example comes from Iris Murdoch:

If I am learning, for instance, Russian, I am confronted by an authoritative structure which commands my respect. The task is difficult and the goal is distant and perhaps never entirely attainable. My work is a progressive revelation of something which exists independently of me. Attention is rewarded by a knowledge of reality. Love of Russian leads me away from myself towards something alien to me, something which my consciousness cannot take over, swallow up, deny or make unreal (2001, p. 87)

In his beautiful and important book, Shop Class as Soulcraft, Matthew B. Crawford quotes this passage and suggests that the same is true of “any hard discipline, whether it be gardening, structural engineering, or Russian,” when “one submits to things that have their own intractable ways. Such hardness is at odds with the ontology of consumerism” (p. 65).

In these cases, a person attends less to another human being than to a complex and integrated object, whether it is a motorcycle, a natural language, or a garden. Duration is a sign of attention. I am not attending to a foreign language unless I commit considerable time to studying it.

The important questions about attention

I have learned from the laboratory research on attention, and I am sure that I would learn more if I read more and would benefit from that study. But this whole literature seems only distantly relevant (if at all) to urgent concerns about attention:

What should we attend to?
Who should we attend to?
What should we do when we notice something about someone else?
Do we attend too much to some things? (Candidates: ourselves, our own pleasures and aversions, trivia, doomscrolling)
Do we attend too little to other things? (Candidates: nature, people who suffer)
Which actions tend to involve worthy attention? (Candidates: skilled manual crafts, “deep reading,” friendship, prayer)
Who and what affects our attention? Is their impact positive or negative?
How can we design contexts in which attention is more often beneficial?

Why the first paradigm came to dominate

Graham Burnett (2026) offers a compelling history of the idea that attention means selecting information to accomplish a task. He mentions that “a huge percentage of the laboratory research on human attention conducted in laboratories in the United States across the watershed decades of the last century was hot-war- and Cold War–directed inquiry into human abilities to track and trigger on a stimulus—work centrally concerned with military operations” (Burnett 2026). Partly due to funding from the Defense Department, but not only for that reason, psychologists used increasingly sophisticated tools for measuring responses to stimuli. I would also emphasize the powerful metaphor of von Neumann-style computers after 1945 and the genuinely fascinating, interdisciplinary field of cybernetics, which peaked in the 1960s. All of these factors help explain why so much research has treated attention as selection of data to process.

Burnett says, “if in the 1950s the entire budget of the U.S. Department of Defense had been rerouted to fund the attentional research of a set of Tibetan monks, we would surely have inherited by 2020 quite a different scientific literature on human attention. Probably less essentially ‘cybernetic,’ and, it stands to reason, less focused on a variety of operationally oriented ‘tasks.'”

As Burnett acknowledges, the fact that any body of research has contingent reasons does not invalidate it. Psychology labs have generated real findings about attention as a solution to cognitive limitations. It’s just that we need to understand other phenomena that also deserve the word “attention” and that seem barely connected to this experimental literature.

Weil’s alternative

Simone Weil (1909-1943) suggests a very different agenda. She begins her short essay (or long aphorism) on “Attention and Will” by announcing, “It is not about understanding new things, but—through patience, effort, and method—coming to understand self-evident truths with one’s whole being.”

As if imagining a psychology lab (although I doubt she had that in mind), she writes, “The will has power only over a few movements of a few muscles, associated with the model of moving nearby objects. I can will to place my hand flat on the table. If inner purity, or inspiration, or truth in thought were necessarily associated with postures of this kind, they could be objects of the will.” But the will is not what concerns Weil. “Attention is an entirely different thing.”

A bit later, she writes,

The poet creates beauty through attention fixed upon reality. The same is true of the act of love. To know that this man—who is hungry and thirsty—truly exists just as I do: that is enough; the rest follows of its own accord. The authentic, pure values ?of truth, beauty, and goodness in human activity arise from one and the same act: a certain application of full attention to the object. The sole aim of education should be to prepare for the possibility of such an act through the exercise of attention.

I conclude with a passage from the same essay:

Solitude: In what, then, does its value lie? For one is in the presence of mere matter (even the sky, the stars, the moon, the blossoming trees)—of things of lesser worth (perhaps) than a human spirit. Its value lies in the superior capacity for attention. If only one could be attentive to the same degree in the presence of a human being…

Sources: Christopher Chabris and Daniel Simons, The Invisible Gorilla: How Our Intuitions Deceive Us (Harmony 2011); John von Neumann, First Draft of a Report on the EDVAC (1945); Iris Murdoch, The Sovereignty of the Good (Routledge Classics, 2001); Matthew B. Crawford, Shop Class as Soulcraft (Penguin 2009); Graham D. Burnett, “Human Attention as a Philosophical Problem: The Question, and the Nature of Questions,” Metaphilosophy 57:1-2 (2026): 3–22; Simone Weil, La pesanteur et la grâce (Librairie Plon, 1947).

why the humanities could never be automated

Leave a reply

Sometimes, we want to answer questions to accomplish practical outcomes. For example, we want to know whether a vaccine works so that we can decide whether to use it. Or we may seek basic insights about viruses so that we can develop vaccines in the first place.

Sometimes, we want to know things because we are simply curious. It is hard to justify a lot of astronomy (for example) on the basis of its practical implications. But we want to understand the universe.

And sometimes, we want to understand people–and perhaps animals–because we are in relationships with them. When you ask friends how they’re doing, your motive may not be to solve a problem, nor mere curiosity, but care. You should give your friend your attention. The benefits are psychological, ethical, or spiritual–a change in one’s mind and in the relationship with the other person.

Friends may tell you things that you should believe for practical reasons or to satisfy pure curiosity. Someone might tell you that a vaccine works, so that you will take it, or that all the planets in our solar system could fit between the earth and the moon, because that’s kind of interesting to know. Paying attention means taking such claims seriously. But the main point of learning what other people think is not to find out what is objectively true; it is to know the other people. In fact, exploring the beliefs and values of a wide range of people can shake our confidence in beliefs, in general, and hence our feelings that we can and must get our own beliefs right.

Attending to a text or an artifact from a distant time or place is a little different from an ordinary conversation. For one thing, you cannot directly benefit long-dead or faraway authors by giving them your attention. Although we have ethical obligations to the dead, the influence is basically one-way. However, reading, listening to music, and viewing art are similar to regular conversations in important ways. They too are practices that develop compassion and reduce our attachment to our own prejudices and concerns.

While Michel de Montaigne’s married friend Diane de Foix was expecting a child, he sent her advice about education. He recommended foreign travel and conversation with peers. A youth should listen and appreciate, he said, not try to form and share beliefs.

And so should adults. Montaigne told de Foix, “In this school of human interaction, I have often observed this vice: instead of getting to know others, we only strive to give ourselves and are more concerned with using our own goods than with acquiring new ones. Silence and modesty are qualities very well suited to conversation.” A passage in the same letter could stand as a justification of Montaigne’s whole way of life:

This vast world, which some think is just one species in a larger genus, is the mirror in which we must look to truly know ourselves. In short, I want [our world] to be my student’s book. The variety of moods, sects, judgments, opinions, laws, and customs teach us to judge our own people soundly and teach our judgment to recognize its imperfection and natural weakness: which is no small apprenticeship.

If the purpose of the humanities–the disciplines that interpret texts and artifacts–is not to determine truth but to attend to other people, then we cannot outsource this thinking to machines or even to other human beings. The point is the experience, not the outcome.

Here is a complication: professional scholars in the humanities do pursue the answers to questions. At one extreme, they may seem much like scientists when they establish the date and provenance of a painting or correct a primary text. At the opposite extreme, they may offer highly creative or even counterintuitive interpretations, but usually they still claim to be telling us something valid about an object in the world. We can ask whether they are right or wrong, persuasive or unconvincing.

I think scholarship is very valuable, but it ultimately contributes to the humanities as experience. The point of watching or reading Shakespeare is to get out of one’s own head by attending to the author and his characters. The point of philological, historical, or interpretive scholarship about Shakespeare is to enrich performances and readings of the works.

Some of the scholarly labor can be assigned to machines. Because I cannot read Pali but I am interested in classical Indian philosophy, I have been very carefully using ClaudeAI to give me dictionary-type definitions of all the Pali words in select passages. This is an algorithmic task. Claude is probably using published Pali-English dictionaries, plus previous translations made by people who used the same dictionaries. A dictionary is also an algorithmic device, a kind of machine that generates a range of words in one language for each word in the other language. I believe that the first dual-language dictionary (Sumerian-Akkadian) was written more than 4,000 years ago.

Thus there is nothing fundamentally new about creating devices that automatically assist readers, listeners, and viewers of human artifacts. In addition to lexicons and dictionaries, we might mention grammars, concordances, indices, card catalogues, provenance lists for artworks, search functions for digital texts, and many other scholarly resources. LLMs can often do these things better.

The difference is that we were never tempted to view the tools as ends in themselves. They were meant to assist a reader in a practice that would enrich that person’s mind. Because the LLM’s speak in the first-person and purport to interpret and explain texts, it is tempting–and I feel this temptation–to imagine that they can do our reading for us. To use a simile that is becoming a cliché, that would be like getting a machine to lift barbells to save us the effort. This would not count as exercise.

Source: Montaigne, vol. 1, essay :26 (“Of the instruction of children”). I follow Screech in interpreting “que les uns multiplient encore comme especes soubs un genre” to mean that our world is one species in a greater genre of worlds.

universities and newsrooms as laboratories or debating societies

Leave a reply

Higher education and the press are important sources of knowledge and insight in modern societies. I think that many people who are interested and concerned about these institutions view them as similar to either 1) labs or 2) debating societies. Both metaphors contain some truth but also mislead.

If you imagine an academic program or a newsroom as similar to a lab, then you will presume that its outputs are information and knowledge. You will probably expect the professionals (professors or reporters) to apply rigorous methods. Any good method counters biases, emotions, and other forms of subjectivity. For example, your own political views should not affect the results of a survey that you conduct if your sample is representative and your statistical techniques are appropriate. In this respect, sampling Americans’ views of Donald Trump is just like taking water samples to measure pH levels. Likewise, your opinions shouldn’t matter if you report on the municipal budget after interviewing a range of insiders and experts.

On this model, you would expect students and novice professionals to learn methods and to be aware of the best supported findings of previous research. Methods and findings should constitute the primary content of education.

When you see a professional consensus about a topic, that is a sign that its methods are working well. Disagreement is problematic, although you can hope that new data or new methods will resolve any temporary dispute.

On the other hand, if you imagine a college as similar to a debating society, then you will think first of a seminar room where there is a free-flowing discussion of a contentious issue (or perhaps a late-night argument in a dorm room). Similarly, you will think first of the op-ed page of a newspaper or a broadcast talk show.

Then you will expect to observe people expressing opinions. Disagreement is desirable–a debate is pointless if everyone agrees–and consensus can be a warning that the whole institution is biased. When someone makes an authoritative claim, along the lines of “We know that X,” you will be quick to suspect them of suppressing alternative views. Your evaluative criteria may include whether the expressed opinions are diverse, whether participants are appropriately open to alternative opinions, whether certain views should be excluded because they are out of bounds, and whether the institution reflects the range of opinions of some appropriate population. (For example, maybe a US broadcast network should present all opinions popular in the US electorate–although that claim is debatable.)

One drawback of the debating society model is that it overlooks the main activities of most professors and reporters: collecting information, applying methods, and reporting results. A session of a college course is much more likely to be spent discussing p-values or prosody than debating politics. In the case of journalism, the number of Americans paid to collect news has fallen by about 77 percent, on a per capita basis, since 1990. There may also be a declining public commitment to academic research across a range of fields.

Also, people who see universities and newsrooms as debating platforms may simply fail to reckon with stubborn information. Sometimes a professional consensus reflects facts, whether we like it or not.

However, if you assume that a university or a newsroom is like a lab, then you will not admit that any question pursued by a journalist or a professor reflects values–beliefs about what is important and why–and assumptions about which methods and sources are legitimate. There may be a neutral way to apply Ordinary Least Squares (OLS) regression to a dataset, but there is no such thing as a neutral dataset. Someone chose to measure certain things because they seemed important.

Instead of being willing to debate and justify your own values and hear critiques of them, you may try to claim that values are irrelevant to your professional work. You will be most comfortable with domains where methods and findings seem relatively uncontroversial, such as the natural sciences and certain kinds of “hard news.” (There either was or was not a fire on Main Street last night).

As topics become controversial, you will become increasingly wary of the observers’ objectivity. For instance, humanities scholars study religion without endorsing specific religions, but you may wonder why something as contestable as a religious belief is a worthy topic, let alone whether an interpretive scholar of religion can be reliable.

For people who see research as value-free science, ethics is unintelligible. It clearly isn’t like a lab science, but if it’s just a matter of opinions, then it isn’t a discipline at all. At best, ethics is a set of legalistic boundaries around the research enterprise, like “Don’t collect data without people’s permission.”

In the modern world, we are confronted with the challenge of navigating both facts and values when the two are deeply connected. We must respect both rigorous methods and free debates. We are trying to grasp truths and honor other people who believe different things. These combinations are difficult.

We might also remember that institutions that are a bit like labs and a bit like debating societies are also other things. Colleges are literal homes for resident students, large-scale employers, institutional investors, landlords, developers, performance venues, and gatekeepers to valuable credentials. Many news agencies are for-profit companies, employers, advertising platforms, and entertainers. Blindness to those realities can make us too comfortable with either model–the lab or the debating society.

the worlds we can lose when intelligence becomes artificial

Leave a reply

In 1958, Hannah Arendt could see where were were headed:

This future man, whom the scientists tell us they will produce in no more than a hundred years, seems to be possessed by a rebellion against human existence as it has been given, a free gift from nowhere (secularly speaking), which he wishes to exchange, as it were, for something he has made himself. …

It would be as though our brain, which constitutes the physical, material condition of our thoughts, were unable to follow what we do, so that from now on we would indeed need artificial machines to do our thinking and speaking. If it should turn out to be true that knowledge (in the modern sense of know-how) and thought have parted company tor good, then we would indeed become the helpless slaves, not so much of our machines as of our know-how, thoughtless creatures at the mercy of every gadget which is technically possible, no matter how murderous it is. (Hannah Arendt, The Human Condition, 1958, p. 3)

What is “human existence as it has been given”?

For most of our history, most human beings have lived with other people whose names they know. They have worked individually and collaboratively with materials in their context to make an environment that I will call a “world.”

A world has these features:

It is imbued with moral significance, because other people have made it, given it meaning, cared about it, and been affected by it. An individual cannot interact with a world without causing good or harm to other people.
It is real, not imaginary, and therefore it is stubborn. It rarely turns out the way we want, but we can learn from experience to work more effectively with it.
The other people involved in any world hold partially conflicting interests and goals and can be stubborn in their own way. Both the materials and the people resist any single will.
Each person has partial and even biased knowledge, beliefs, and feelings about the world. But their varied ideas can accumulate as they express them and record them. Each person can therefore explore not only a world but the accumulated human experience of that world.
Because we must act in the company of other people and learn by acting, our “thinking and speaking” are closely connected.
Because our deepest concerns (moral, spiritual, and otherwise) relate to the world that we shape with our minds and hands, our “thought” is also connected to our “know-how.”
Each world typically predates each human being and survives the person’s death, yet each person can affect it. In fact, the birth of any human being automatically changes the world, if for no other reason than a birth turns people into parents, siblings, and other kinds of relatives.
There is not one world but many human worlds. But worlds can interact to various degrees without becoming subsumed into one bigger world.

Why it is good to live in a world

It is not obvious that living in this kind of world is the best imaginable form of life. Most people have envisioned heaven or a political utopia differently. (For instance, in an ideal world, the other people usually become less stubborn!) But I could make three arguments in favor of living in a world like this.

First, it seems plausible that homo sapiens evolved for such a life. Our brains, senses, and bodies are equipped to navigate it.

For instance, newborn infants already recognize faces, which are designed to communicate information and emotions. And our languages and cultures have accumulated deep resources for sharing a world with finite other human beings. The Proto-Indo-European language already used first-, second-, and third-person verbs and indicative, imperative, and subjunctive moods to make distinctions that are useful for group discussions about a common world. Thus a world is arguably our habitat.

Second, the combination of agency and humility seems morally compelling. It is fitting that we can affect our environment but not do just anything we individually want with it. And we should see our context as imbued with moral significance.

Third, navigating a world is a way for creatures like us to achieve comprehension, to make sense of matters. As Arendt writes:

There may be truths beyond speech, and they may be of great relevance to man in the
singular, that is, to man in so far as he is not a political being, whatever else he may be. Men in the plural, that is, men in so far as they live and move and act in this world, can experience meaningfulness only because they can talk with and make sense to each other and to themselves (1958, p. 4).

Threats to human worlds

Each human world has always been fragile, subject to destruction if invaders arrive, a plague strikes, or the community breaks down.

In addition, tyrants threaten any shared world because they can turn individuals into means to their solo ends.

Mass society puts each world at risk by bringing us into relationships with millions of others, whose names we will never learn. And mass economic exploitation makes matters worse. In Origins of Totalitarianism, Arendt says, “loneliness, on the experience of not belonging to the world at all, … is among the most radical and desperate experiences of man. [It is] is closely connected with uprootedness and superfluousness which have been the curse of modern masses since the beginning of the industrial revolution and have become acute with the rise of imperialism at the end of the last century and the break-down of political institutions and social traditions in our own time.”

When history seems to move quickly and beyond anyone’s control, humans cease to feel that they are agents in any recognizable world.

Ideology can be defined as any system of thought that substitutes core assumptions for actual engagement with other people in a common world.

Finally, although media can enrich any given world, it can also disrupt it. Imagine people sitting alone or in passive company before a TV screen that tells them about gruesome crimes. Their actual world may be safe, or less dangerous than it was in the past, but the mediated world is cruel.

New threats in the age of AI

This theoretical framework comes from Arendt, who drew on Heidegger’s fundamental insight that the human form of being (Dasein) is always “‘in’ the world in the sense that it deals with entities encountered within-the-world, and does so concernfully and with familiarity” (Being and Time, H105, trans. by Macquarrie & Robinson). Arendt makes Heidegger’s theory political and republican by emphasizing that people can talk and decide what to do with their worlds.

I have sketched this view to help make sense of a new phenomenon: intelligence that is artificial (AI). But Arendt already feared that we might “need artificial machines to do our thinking and speaking.”

When a person expresses a view, the content of what they say helps us to understand the world that the person inhabits. Even when people are flat-out wrong, the fact that they err or lie is part of our reality. In addition, a human view comes from a creature that can suffer. As such, it makes a claim on our compassion. In short, we attend not only to the content of the statement but also to the person who expressed it.

In contrast, when a large language model (LLM) answers a query (typically in the first-person singular and with emotive language like “I will be glad to …”), it does not reflect any particular perspective, nor does it come from a body that is capable of suffering. It just pretends to be a fellow participant in our world. We can attend to the words but not to the speaker.

Walter Cronkite was not really a visitor to Americans’ living rooms in 1970. He just appeared on TV screens. But he was a real person who could be assessed as such. An LLM is qualitatively different.

An LLM can be just another tool or resource, like a Heidegger’s hammer or perhaps like a library. I have collaborated with teams of Tufts engineering students to build the Civic Helpdesk and other applications of AI that are not yet publicly available. Working with them to fine-tune instructions or to design a user interface feels very much like collaborative work in a shared world. Note that I naturally said we “built” these tools, because the work feels roughly like building a shed, or perhaps an organization.

I have also developed what I think is a fairly tight practice of asking Claude about the Sanskrit and Pali original words in texts that I can only read in translation. This feels like a modest expansion of my inner life, if not a contribution to any shared world. (By the way, Claude is probably pulling these definitions from a finite set of published lexicons that have human authors.)

On the other hand, as Pope Leo notes in Magnifica humanitas, “current AI systems are more ‘cultivated’ than ‘built,’ for developers do not directly design every detail, but instead create a framework within which the intelligence ‘grows.’ As a result, fundamental scientific aspects — such as the internal representations and computational processes of these systems — remain, at present, unknown.” This sounds more like Arendt’s nightmare of a time when our thoughts cannot grasp what we have done.

The deepest concern is that we have developed biologically and culturally to flourish in what Arendt would call a world, but an individual who uses AI is no longer there.

Peter Levine

A Blog for Civic Renewal

Category Archives: science, technology and society

what it would take for the public to trust science

paradigms of attention

why the humanities could never be automated

universities and newsrooms as laboratories or debating societies

the worlds we can lose when intelligence becomes artificial