On complex political issues, most Americans take shortcuts to simplicity

A novel study by UC Berkeley scholars finds that when Americans are presented with complex decisions on immigration, climate change or political candidates, they often avoid difficult reasoning

University of California - Berkeley

Facebook

XLinkedInWeChatBlueskyMessageWhatsAppEmail

A core assumption in any democracy is that citizens with access to information will make reasoned decisions on political candidates and policy issues. But a new study by UC Berkeley social scientists finds that Americans usually don’t consider the full scope of difficult issues, instead basing their opinions on mental shortcuts that sidestep complexity.

The paper offers striking insight into the differences in cognitive reasoning patterns among Americans. While some people in the study did closely analyze a range of issues before coming to an opinion, others did a quick back-of-the-napkin tally of plusses and minuses, or even more commonly based their opinions on just one factor, the authors found.

“When people make complicated choices with many considerations — which political candidate to support, which immigrant they would rather admit to the country, which climate policy they prefer — we tend to assume they're weighing all the factors,” said Berkeley political scientist Kirk Bansak. But, he added, “we find that most of them aren't. Most commonly, people rely on a mental shortcut, effectively deciding based on the single factor they care about most and largely ignoring the rest.”

For example, when asked to choose between two fictional U.S. Senate candidates based on age, gender, race, education, military experience and political experience, many of the study participants would make their decision based on just one of those factors.

 That doesn’t mean that people are irrational, said co-author Nidhi Banavar, a Berkeley cognitive scientist. “The kind of rationality that we define in paper is quite different from what many humans actually do in real life,” she explained. “We have these amazing brains, but they're not the infinite computational machines that would be necessary in order to do the kind of ‘rational’ we imagine in our paper.”

 Still, the findings have clear implications for politics and political communication. For example, Banavar said, if a leader knows that public opinion on climate change is dominated by pocketbook concerns to the exclusion of other issues, then that single factor might weigh heavily in political advertising. Complex science issues, meanwhile, might be downplayed.

 The paper, “Classifying Strategy Use in Multi-Attribute Subjective Choice: Application to Conjoint Experiments in Political Science,” was published earlier this week in Proceedings of the National Academy of Sciences. Bansak is an associate professor of political science; Banavar is a postdoctoral researcher in political science.

 The foundational ideals of American democracy grew from the Enlightenment era some three centuries ago. Central to that framework was a belief that humans, being rational, are capable of governing themselves — and justified in demanding political power. A core concept has been the marketplace of ideas, a constant competition between political and policy goals and plans that would inform citizens and help them to govern themselves.

 While many eras in U.S. history feature episodes in which reason faltered, modern social scientists have detailed the ways in which cognitive bias, rational ignorance and other behaviors short-circuit reasoning. Still, the hopeful assumption about human reasoning persists.

Deep patterns in decision-making

Social scientists have long studied human decision-making at the aggregate level. Banavar and Bansak set out to build a new kind of cognitive model, a complex tool that can enable researchers to see an individual’s decision-making mid-process, in detail.

 The model doesn’t give them a view into the minds of their research subjects, but the researchers can study hundreds or thousands of decisions by an individual; from these patterns, they can perceive the values and reasoning processes that shape their decisions.

 In all, nearly 3,800 people were involved in the study. Each participant sat at a computer screen, and the computer generated a random set of choices in three areas: which migrants they would rather admit to the United States, which Senate primary candidate they would rather vote for, and which climate policy package they would prefer the United States to adopt.

 Consider the basic question comparing two hypothetical migrants to the U.S. Their profiles appeared side by side, each described by six factors: region of origin; education level; gender; age; English proficiency; and whether they were seeking economic opportunity or asylum. The research subject would have to choose which of the migrants they preferred to admit to the U.S. 

The computer randomly presented 30 different sets of hypothetical migrants. If a participant consistently chose the younger migrant or the migrant with greater English skills, for example, the researchers could conclude that that was the most important factor for that participant.

The same decision framework was presented for comparing potential climate change policies and fictional U.S. Senate candidates. 

The research did yield some insights about basic priorities: Participants preferred younger political candidates. They preferred climate change policy that minimized economic impact on their household. They preferred immigrants who were seeking asylum as opposed to seeking economic opportunity, and those who were below retirement age. 

But that wasn’t the primary purpose of the study. Banavar and Bansak were less interested in what participants decided, and more interested in how they decided.

Thoughtful reflection or hot takes?

The results were striking: In 68% of the cases, study participants used shortcuts and did not deeply assess all six factors when choosing between hypothetical Senate candidates, migrants or climate policies.

Nearly 54% showed a pattern of making choices dominated by one of the six factors. Another 14% took a different shortcut: Given six factors for each candidate or policy, they counted the total number of factors that they preferred for each, then chose the one containing the greater number. 

How many used a comprehensive evaluation of all six factors — a rational assessment, without any shortcuts? Just under 30%.

Interestingly, Banavar said, a majority of participants did not consistently stay with the same strategy. Some might rely on just one factor in choosing a climate policy, while more fully weighing all the factors in deciding which migrant to admit.

“You can't put people into buckets to say, ‘Okay, this is someone who's actually only going to care about only one thing,’” she said. “People are actually relating differently to each of these different domains and choosing to use different levels of effort, for want of a better term.”

To understand why people shift strategies will require further research, she said.

The study comes with some important caveats: If a participant favored female candidates, for example, or migrants who were seeking political asylum, and if both choices had those attributes, the participants had to consider other factors. That reduced, but did not eliminate, the pattern of making a choice based on one factor.

But when the researchers asked the participants to rate each profile before choosing one of them, the participants were more inclined to consider all factors in making their choices.

Further, the researchers allowed for only two attributes in their comparisons — male or female, for example, white or Black, or aged in their 50s or 70s. That’s not ideal, Banavar said, but to include multiple races, for example, is too complex for the cognitive model, and would dilute the patterns in the data. And in the choice of Senate candidates, the study did not include the fictional candidates’ political parties. That question could provoke such strong, reflexive reactions that it would cloud the findings, she said.

Implications for real-world politics

The authors say their research has important implications not just for cognitive research, but for real-world, on-the-ground U.S. political processes.

Often, Bansak said, our understanding of public opinion is based on surveys that ask people to choose between complex options defined by many factors. The new research shows analysts that individuals often base their opinions on a single factor, and that can create a clearer picture of what people want in public policy, and why they want it.

“This bears directly on how we interpret the kind of public-opinion and preference data that routinely informs policy debates and news coverage,” he said.

And that understanding, Banavar said, can shape the policy process — and political communication. If leaders present issues in layers of complexity, “how much are people going to read or listen to that?” she asked. “What is the thing that's really important to them, given a particular issue? That's completely lost when you don't think about the decision-making process.”

---

DOI

10.1073/pnas.2613113123

Method of Research

Computational simulation/modeling

Subject of Research

People

Article Title

Classifying strategy use in multiattribute subjective choice: Application to conjoint experiments in political science

 

The radical ecology of Rosa Luxemburg

By Paul Le Blanc

Published 24 September, 2026

The following is a review by Paul Le Blanc of Rosa Luxemburg’s Herbarium: Radical Ecology and the Global Plantation, edited by Claudia Horn. New York: OR Books, 2026. 256 pages. $30 US/ £23 UK. This review is simultaneously published on LINKS and Communis.

Rosa Luxemburg’s Herbarium: Radical Ecology and the Global Plantation is a handsomely crafted and inspiring volume to be savored and shared, thanks to the uniquely exquisite sensibilities of Rosa Luxemburg and the splendid labors of the book’s editor, Claudia Horn. Richly illustrated by Luxemburg herself — with beautiful reproductions from among the many hundreds of samples in the 17 notebooks that contain Luxemburg’s drawings of vegetation and the pressed plants themselves, enhanced with numerous marginal notes by her in German — the book also provides well-chosen slices from correspondence with some of her closest friends, along with insightful and informative commentary from the editor. Revealed to us are little known dimensions of Luxemburg the outstanding thinker and revolutionary, but also the generously gifted human being.

Qualities of thought and revolutionary practice

Luxemburg wanted to replace the established order with a genuinely democratic and socialist future. The old order in Germany had taken the form of voracious capitalism integrated with a reactionary elite of land-based aristocrats. Much has changed in the world since, and yet Luxemburg’s perspectives have never stopped resonating in countries throughout the world to the present moment.

For her revolutionary commitments, she was denounced as “bloody Rosa”. Many (including moderate socialists) agreed with the highly esteemed centrist-liberal/conservative sociologist Max Weber. He was fiercely critical of the Spartacus League, which Luxemburg led with her comrade Karl Liebknecht in 1919. Several days before they were brutally killed in the so-called “Spartacist uprising,” Weber said that “Liebknecht belongs in a lunatic asylum, and Rosa Luxemburg in a zoo.” While he had not advocated their murder, it seems not to have shocked him. “Liebknecht was undoubtedly an honest man,” Weber commented, once the deed was done. “He called on the street to fight — the street killed him.” Much the same could have been said of Luxemburg. Of course, they were not killed by “the street,” but by a right-wing paramilitary, the Freikorps, commanded by officers from Germany’s reactionary elite, with tacit support from more moderate “law and order” elements.1

The personal qualities of “bloody Rosa” — who spent years in prison for her beliefs — do not correspond, though, to Weber’s anti-revolutionary caricature. Horn cites the testimony of one of the many prison staff members who befriended her (in this case a non-commissioned officer, Arthur Gertel, who was tasked with accompanying her on walks that she was permitted to take):

She was an intellectual genius yet full of kindness and compassion for all beings, human or animal… And yet the newspapers called her “Bloody Rosa.” The sharpness and refusal to make fundamental compromises that characterized Rosa Luxemburg’s political speeches and writings undoubtedly stemmed from her outrage at all injustice.2

It is the crescendo of injustice and horrific violence — inherent in capitalism, imperialism and global war — that would finally bring an end to Luxemburg’s life in 1919. That is not the focus of this volume. What Horn presents to us, however, transcends the violence in a way that refuses to push aside Luxemburg’s determined struggle to replace it with something better. As she emphasizes, “Luxemburg’s expansive sensitivity to human and non-human suffering drove her unswerving, unfailing, and uncompromising commitment to the cause of world revolution.”3

Enlightenment, Romanticism and the natural world

The great intellectual-cultural streams of the Enlightenment and Romanticism blend in this remarkable woman, with — at the very same time — rock-hard yet supple Marxist convictions that had become inseparable from a passionate yet very conscious connection to the natural world. Her strong bond with blue-feathered titmice is especially evident. She would feed them between the bars of her prison window and copy their bird song. In a letter to Mathilde Jacob, Luxemburg boasted that she could imitate them so well “that they all immediately come running,” adding: “In spite of the snow and frost and loneliness, we believe — the titmice and I — in the coming of spring!” In another letter to Sophie Liebknecht, she confessed that “sometimes, it seems to me that I am not really a human being at all, but rather a bird or a beast in human form.” In yet another letter, she proudly confided to Luise Kautsky:

The great titmice are in loyal attendance in front of my window; they already know my voice exactly; and it seems they like it when I sing. Recently I sang the Countess’s aria from Figaro, about six of them were perched there on a bush in front of the window and listened without moving all the way to the end; it was a very funny sight to see.4

Little wonder that she fantasized: “on my grave, as in my life, there will be no pompous phrases. Only two syllables will be allowed to appear on my gravestone: ‘Tsvee-tsvee’” (her transliteration of the titmouse call that she had mastered). While expressing the fervent hope that her death would come in the struggle for the socialist cause, she confessed “my innermost self belongs more to my titmice than to the ‘comrades’.”5

In fact, her expansive temperament embraced the whole of the natural world. “Inwardly,” she wrote, “I feel so much more at home in a plot of garden … and still more in the meadows when the grass is humming with bees, than at one of our party congresses.” As this little volume reveals, Luxemburg shared the passionate engagement of the bees and other creatures with the flourishing and sometimes struggling plants of all kinds — which delighted and fascinated her. Yet she insisted that nature was not, for her, “a restful refuge,” noting that “in nature too, at every step, I find so much that is cruel that I suffer very much.”6 The point was to be vibrantly aware, conscious of, and in tune with the wondrous reality of life — deeply in touch with herself and the vast and complex world around her.

In more than one way, this informed her political understanding and orientation. Luxemburg tirelessly analyzed the causes and impacts of the catastrophic horrors of World War I, with the conviction that “by its nature, socialism cannot be imposed” — that the key to reaching it could only be found through “uninhibited, effervescing life,” which meant that “experience alone is capable of making corrections and opening up new paths.” While the failure of the international socialist movement to stay true to its principles in resisting the global imperialist war brought a devastating “misfortune for humanity,” she insisted, “socialism will be lost only if the international proletariat refuses to learn from it.”7 Revolutionary consciousness and insurgency cannot be manufactured from the top-down. It must flow from actual experience as a collective process.

Capitalism and metabolic rift

Even more than was the case with the cataclysms of global war in 1914-18 and 1939-45, all human and non-human life on our planet is threatened by unfolding and intensifying ecological disasters. Just as imperialism and the global wars it generated were not simply the fruit of mistaken policy decisions — according to Luxemburg’s penetrating analyses — so is the environmental crisis not the outcome of policies that can simply be reversed. Both flow inexorably from the accumulation process at the heart of the capitalist economic order in which we live. Horn notes the same point when she writes, “profit seeking capitalist behavior causes various kinds of economic imbalance.”8

This imparts a special urgency to aspects of Luxemburg’s application of ecological sensibilities to her development of Marxist analysis. It is related to the notion of “metabolic rift,” recently highlighted in significant studies by such “eco-Marxists” as John Bellamy Foster and Ian Angus.9 Foster and Angus, following Marx — as Horn puts it — use the metaphor of “metabolism” to describe “the complex, interdependent process linking human society to nature.” According to Horn, “all economies transform — metabolize — physical materials such as wood, water, and soil into goods for use or exchange.” But “if producers do not exploit those natural resources in a sustainable way, both the resources and the economic system they support may be exhausted.” 

Drawing from Luxemburg’s classic The Accumulation of Capital (1913), Horn presents the view that “capitalism was not a closed system but rather operated within and depended upon a wider, non-capitalist ‘milieu’.” In fact, “capital accumulation had required the continuous absorption and transformation of resources — both human labor power and natural materials like wood or minerals — that originated outside capitalist societies.” Capitalism existed “not as a stable equilibrium but as a ceaseless, violent churn.” This was stressed in Luxemburg’s theory of imperialism as a perpetual invasion and exploitation of non-capitalist portions of our planet. 

In Luxemburg’s words, “capitalism feeds on the ruins of such organizations, and, although this non-capitalist milieu is indispensable for accumulation, the latter proceeds, at the cost of this medium, by eating it up.” Horn draws out this point: “Eating up the conditions for its survival, ‘capitalism prepares its own downfall under ever more violent contortions and convulsions.’” Nonetheless, as Horn observes later, “the capitalist system continues to displace the real costs of production, such as environmental destruction, outside the realm of capital accumulation, to be borne off-books by exploited people, non-human animals, and future generations.”10

Horn strongly suggests that Luxemburg’s particular analysis of imperialism suffered from its own flaws, to which she gives serious attention, but insists on the relevance of much of what she had to say about the capital accumulation process:

If history has not borne out Luxemburg’s argument that capitalism depends upon both exploiting and incorporating the non-capitalist world, such that it must inevitably exhaust itself, still, the enduring relevance of her analysis lies in its global scope; its sensitivity to the physical predicates of economic growth; its focus on the boundaries that capital constantly transgresses, polices, and redraws; and its attentiveness to the broad coalitions these trespasses and depredations could bring together in resistance.

The qualities of this remarkable book certainly do justice to the multifaceted person who was Luxemburg, but they also speak to us on a variety of levels about the meaning of life and about what must be done in the present moment.

• 1

  Alexander Behrens, review of The Young Max Weber and German Social Democracy by Victor Strazzeri (Brill 2022/Haymarket, 2023) in The Progressive Post, July 16, 2024, https://feps-europe.eu/inspiration/the-young-max-weber-and-german-social-democracy/.

• 2

  Claudia Horn, ed., Rosa Luxemburg’s Herbarium, 14.

• 3

  Ibid., 19.

• 4

  Ibid., 15-16, 48, 136-137. Blue tufted titmice: https://www.youtube.com/watch?v=zcQOEMxYUMU; Mozart aria: https://www.youtube.com/watch?v=HaZTuRBFtDk.

• 5

  Rosa Luxemburg’s Herbarium, 15-16.

• 6

  Ibid., 16.

• 7

  Ibid., 16, 13.

• 8

  Ibid., 52.

• 9

  See, for example, John Bellamy Foster, Marx’s Ecology: Materialism and Nature (New York: Monthly Review Press, 2000) and Ian Angus, Metabolic Rifts: Capitalism's Assault on the Earth System (New York: Monthly Review Press, 2026).

• 10

  Rosa Luxemburg’s Herbarium, 51, 52, 55-7, 58, 211. 

Op-Ed

Big Tech’s Debate Over “Open” and “Safe” AI Leaves Corporate Power Unquestioned


Two seemingly contradictory letters emerging from the AI industry point toward the same future.

By Tim Scott ,

Truthout

September 1, 2026

Anthropic CEO Dario Amodei looks on during a meeting with France's President Emmanuel Macron on the sidelines of the AI Impact Summit in New Delhi, India, on February 19, 2026.Ludovic MARIN / AFP via Getty Images

On July 24, 2026, an initial coalition of 25 technology companies and organizations, including NVIDIA, Microsoft, and Meta, released the Open Weights and American AI Leadership letter, calling for wider adoption of AI models that institutions can download and modify. By August 3, more than 270 companies and organizations had signed, spanning Big Tech, venture capital, investment management, and the defense and security industries. Four days after Open Weights’ release, another letter, Pacing the Frontier, organized by employees of frontier AI companies with support from two AI-policy nonprofits, drew more than 1,300 signatures. It warned that leading companies may be close to automating AI research, accelerating development beyond society’s ability to understand or control the resulting AI systems.

The letters seem to argue for opposite futures. One presents wider access and fewer restrictions as the route to security and prosperity. The other asks the U.S. government to support an international effort to slow frontier development before competitive pressures make coordinated restraint impossible.

But they address different levels of the industry. Open Weights seeks to deepen AI’s already pervasive presence by making customizable models easier to embed throughout the economy and public life. Pacing the Frontier proposes government-backed management of frontier AI, the industry’s term for its most advanced and broadly capable models. That framework could make the small circle of companies building these systems indispensable to defining the dangers and writing the rules meant to restrain them.

Read together, the letters point toward a two-tier arrangement. AI becomes entrenched across society while ownership of its most powerful systems, infrastructure, and governance remains concentrated at the top. Their connection requires no secret agreement. It lies in what both take for granted: AI development will continue, public institutions will finance and adopt it, and private ownership and corporate power will remain intact.

The letters leave out harder questions: Who will own AI and decide its direction? How will it be used in surveillance and war, and how will it reshape work and deepen inequality? Whose lives and futures will be valued, and does the public have any right to limit or refuse a technology presented as inevitable? Beneath the language of openness and safety lie deeper a general agreement over corporate power, military dominance, elite rule, and the authoritarian currents shaping Silicon Valley’s vision of the future.

Open Weights Are Not Open Source

An AI model is the underlying system that generates answers, predictions, images, or other outputs. Its “weights” are the billions of numerical settings developed from patterns in training data. Releasing those weights allows organizations with computing power and expertise to download, run, and modify the model rather than use it only through a company-controlled service such as ChatGPT, Gemini, or Claude.

Researchers can examine these models, test their weaknesses, and build specialized tools, while organizations can run them locally rather than send sensitive information to a commercial service. Yet proprietary, company-controlled services such as ChatGPT, Gemini, and Claude are not inherently safer: their operators can keep flaws from public scrutiny, monitor how people use them, impose restrictions, raise prices, or withdraw access.

But releasing a model’s weights does not by itself make the AI system open source. Under the Open Source Initiative’s definition, open-source AI must provide the necessary materials and the legal freedom to use, study, modify, and share the system for any purpose. Those materials include the model’s parameters, the complete code used to train and run it, and detailed information about its training data and methods. Most open-weight releases, by contrast, provide the finished numerical settings and enough supporting files to run or adapt the model, but not the complete training code or the required details about its training data and methods. The training data itself and much of the development process often remain proprietary or unavailable. Beyond the definition’s formal requirements, the labor arrangements and corporate decisions behind these systems also remain undisclosed. Users of open-weight models can therefore operate and adapt them without gaining a full account of how they were built.

The Open Weights letter borrows the democratic reputation of open source to present open-weight access as a path to competition, independence, and shared prosperity. Yet an open-weight release gives the public access to a corporate product without transferring ownership or giving workers, affected communities, and the broader public meaningful power to govern the data, computing infrastructure, labor conditions, and deployment decisions behind it. A genuinely open-source release would provide broader materials and legal freedoms, but even that would not place the resources, infrastructure, or institutional uses of AI under public, community, or worker governance. In either case, access remains unequal. Individuals may run smaller models, but corporations and agencies possess the data centers, private datasets, institutional records, and specialists needed to customize them, connect them to institutional data systems, and deploy them at scale. An open-weight model that supports independent research can therefore also expand institutional surveillance and coercion by enabling governments and corporations to monitor and classify people, inform or automate consequential decisions, and control access to services.

That unequal distribution of power is only one danger. Open weights also create an irreversible loss-of-control problem. Once copied, a model cannot be fully retrieved, monitored, or patched across all distributed copies, nor can its safeguards be guaranteed to remain intact. Those with sufficient technical expertise and computing resources can modify open-weight models, remove their refusal mechanisms, introduce discriminatory behavior or hidden backdoors, and redistribute the results. Modified models could facilitate cyberattacks, fraud, impersonation, targeted disinformation, nonconsensual sexual imagery, harassment, or guidance involving weapons, abuse, and self-harm. Connected to institutional records, they could expose private information, intensify surveillance, discriminate, restrict services, or help target communities. Responsibility then fragments as developers, vendors, and institutions blame one another.

Even as the weights circulate and open the door to still more harmful uses, the infrastructure remains firmly closed. AI depends on specialized chips, enormous data centers, electricity, water, and cloud platforms. Open weights may reduce an institution’s dependence on a particular model provider, but training, adapting, and deploying AI at scale still depend on highly concentrated infrastructure markets, led by NVIDIA in advanced AI accelerators and by Amazon Web Services, Microsoft Azure, and Google Cloud in cloud computing. All four companies now appear among the letter’s signatories. The model may be downloadable, but control over the infrastructure surrounding it remains concentrated in a small group of corporations.

For researchers and organizations in the Global South, open weights may provide some genuine independence from closed commercial services and make it easier to adapt models to local languages and needs. But that limited autonomy does not alter the global labor system through which AI is produced, requiring a vast workforce largely across the Global South that gathers, labels, cleans, translates, moderates, and evaluates data and model outputs. These workers are often monitored and managed algorithmically, bound by nondisclosure agreements, and required to endure prolonged exposure to traumatic material. Releasing a model’s weights reveals none of these labor conditions and gives the workers neither an ownership stake nor decision-making power over the systems their labor helped create.

Open Weights asks policymakers to expand computing access, fund datasets and testing tools, avoid restrictions that might slow development, and spread AI throughout the economy. This is a familiar Big Tech strategy: use public resources to build markets for privately controlled technology and make everyday life dependent on products the public neither owns nor governs. The costs are distributed across workers and the public, while private companies retain control over much of the infrastructure, intellectual property, and resulting wealth.
Pacing the Frontier — But Under Whose Authority?

Building frontier AI requires extraordinary amounts of computing power, energy, data, and capital, largely confining its development to a small number of Big Tech corporations, heavily financed AI laboratories, and state-backed institutions. Pacing the Frontier warns that if AI systems begin automating the research and engineering needed to create more powerful successors, development could outpace researchers’ and governments’ ability to test, understand, or control them. These systems could facilitate cyberattacks, weapons development or deployment, mass manipulation, and other catastrophic harms before effective safeguards are in place. Because no company or country wants to surrender its competitive position while others continue racing ahead, the statement calls on the U.S. government to support an international effort to develop technical and governance mechanisms for deliberately pacing frontier development.

Many of the letter’s signatories may sincerely seek to prevent catastrophe, and coordinated restraint may indeed be necessary. Yet the letter calls for no immediate pause and leaves the decisive questions unanswered: Who decides when development must slow? What capability threshold or kind of security breach would trigger intervention? Who oversees the companies, and what prevents the largest laboratories from shaping rules that entrench their dominance by imposing barriers smaller competitors cannot meet?

Many of the signatories work for companies that would be subject to any resulting rules, yet these same companies possess enormous power to shape those rules. These firms are private centers of power and rank among the world’s wealthiest and most politically connected institutions. Collectively, they own or control much of the frontier-model, cloud, and data-center infrastructure, as well as the proprietary information and technical expertise on which regulators depend. Their wealth funds lobbying, research partnerships, political access, and policy campaigns. Through search engines, social platforms, AI assistants, media partnerships, and sponsored research, they also shape public perceptions of AI: what it promises, which harms receive attention, and which alternatives appear possible.

Yet this power is subject to little democratic accountability. Major decisions are made without public scrutiny, while the people and communities affected by these systems have little power to inspect them, challenge their decisions, or refuse their use. The companies can help define what constitutes a danger, which responses are acceptable, and which restrictions remain outside serious debate. Those creating the risks thereby become indispensable government partners and coauthors of the rules meant to restrain them.

The emphasis on future catastrophe can push present harms out of view. AI is already intensifying racist systems of mass surveillance, policing, detention, and deportation; supporting military targeting and the rapid development of increasingly autonomous weapons systems; automating work and displacing workers; deepening data extraction; and accelerating environmental destruction. A Brennan Center report details racial disparities in facial recognition and notes that Black people account for at least 8 of 10 known wrongful arrests based on faulty matches. The Department of Homeland Security AI inventory documents ICE’s use of facial recognition. An ACLU investigation traces Palantir’s role in ICE’s detention and deportation infrastructure. A Government Accountability Office review describes the Pentagon’s Maven program using machine learning to identify potential targets from geospatial imagery. The AI Now Institute’s 2025 Landscape Report documents AI’s energy and water demands, worker displacement and exploitation, workplace surveillance, deepening inequality, and risks to critical infrastructure and national security. Debate over hypothetical loss of control can obscure the harms already imposed on people by systems functioning as designed.

Pacing the Frontier does little to challenge the structures producing these harms. Its limits are especially clear on labor. A genuine slowdown could reduce the pace of worker displacement, resource extraction, and demand for exploitative or psychologically harmful data work, but it would not by itself change the ownership structures, working conditions, or power relations governing AI production. The letter does not include these present harms in its case for restraint. It defines danger primarily through frontier capabilities, geopolitical competition, and possible future catastrophe, not through low wages, algorithmic management, psychological injury, labor displacement, or the transnational contracting chains already sustaining AI development.

That narrow conception of danger also shapes the safety framework envisioned by Pacing the Frontier. Evaluating and stress-testing models, reviewing content, and monitoring systems after deployment would still require human labor. Given the tech industry’s entrenched labor practices, this framework would almost certainly continue to rely on workers who toil in the shadows. The letter calls for none of the protections or decision-making power these workers would need: labor standards, transparency about who performs the work and under what conditions, collective bargaining rights, protections against workplace surveillance, or a meaningful role for data workers in determining what their labor is used to produce.

The concentration of authority in laboratories and government experts can fortify the companies already at the top. Frontier regulation may reduce genuine dangers, but expensive testing, security, and reporting requirements can exclude universities, public-interest researchers, and smaller developers while the largest firms absorb the costs. Because governments depend on corporate expertise, oversight can become a wall shielding Big Tech from public accountability and leaving the labor regime on which it depends untouched.

On labor, then, the letters are less opposed than they appear. Open Weights would diffuse models whose dependence on human labor remains concealed; Pacing the Frontier would subject the most powerful models to expert oversight without democratizing the labor relations or restructuring the supply chains behind them.
The Ideology Behind the Two-Tier System

Both letters treat AI’s advance as inevitable, leaving the public to adapt and government to manage its deployment. That technological determinism merges with Silicon Valley solutionism: questions about ownership, labor, surveillance, military power, and the right to refuse AI shrink into matters of design and oversight. Within that shared premise, Open Weights gives institutions greater freedom to deploy AI, while Pacing the Frontier places corporate laboratories, experts, and officials at the center of determining whether and when AI development poses an unacceptable danger. The public remains a population to manage rather than a political force empowered to decide whether and how the technology should be developed and used.

The division between the letters is not ideological alone. It also corresponds, although imperfectly, to different positions and material interests within the AI economy. The largely industry-based coalition behind Open Weights spans chipmakers, cloud providers, model developers, software companies, security and military contractors, and venture capital firms. Pacing the Frontier emerges from a different institutional position. It was signed in individual capacities by employees, researchers, and executives of frontier AI companies, including some of the industry’s most influential figures, rather than formally endorsed by the companies themselves. Their concerns cannot simply be reduced to their employers’ economic interests. Yet their place within the industry matters: they work inside the small group of elite laboratories where computing power, investment, proprietary knowledge, and technical authority are most heavily concentrated.

The letters, however, should not be treated as expressions of completely separate blocs of capital. Google, Meta, OpenAI, and other companies appear in the Open Weights coalition, while employees and executives of the same firms signed Pacing the Frontier. A single corporation can benefit from broad diffusion across the AI market and concentrated control at the frontier: it can encourage widespread adoption, sell the infrastructure and services that adoption requires, and retain control over its most advanced systems. What appears to be a conflict among different sections of AI capital can therefore also reflect competing strategies within the same corporations and investment networks. This is primarily a dispute within AI’s corporate and technical elite over the terms of expansion, not a challenge to corporate ownership or concentrated control.

Open Weights echoes themes found in Marc Andreessen’s Techno-Optimist Manifesto, effective accelerationism, and techno-libertarianism. These currents treat technological expansion as a social good, private markets as engines of progress, and regulation as an obstacle. Pacing the Frontier departs from techno-libertarianism’s hostility to government restraint, but it does not overturn the distribution of power that techno-libertarianism legitimates. Frontier systems remain privately owned, development remains organized through corporate competition, and technical and corporate elites retain a privileged role in defining danger. The letter therefore represents less a rejection of techno-libertarianism than a technocratic effort to contain the dangers created by unrestrained corporate competition while leaving its underlying structures of ownership and control intact. At the same time, its emphasis on possible future catastrophe and expert-led restraint more closely resembles longtermism, which can assign greater moral importance to hypothetical future populations than to people suffering now, and singularitarianism, which anticipates that superhuman AI will transform society beyond recognition. These currents overlap with transhumanism, which envisions using technology to enhance selected human capacities and ultimately create supposedly superior “posthuman” forms of life.

Despite their differences, these ideological currents converge around the pursuit of ever-more-powerful AI. Accelerationists seek more computing power, faster deployment, and fewer restraints, convinced that technological growth will produce abundance and solve social problems. Longtermist safety frameworks may urge greater control over the pace of development, but they often leave that underlying project and its private corporate ownership and direction unquestioned. In the framework developed by Timnit Gebru and Émile P. Torres, these visions share a tendency to centralize power, elevate technical elites as humanity’s guardians, and invoke “safety” or universal benefit while pushing out of view the exploitation, discrimination, and environmental destruction already affecting dispossessed peoples.

They trace transhumanism, singularitarianism, longtermism, and related ideologies to Anglo-American eugenics. In their account, transhumanism recast improving the “human stock” as creating enhanced or “posthuman” beings. The language shifted from compulsory breeding to innovation and personal choice, but the sorting logic endured: ideas about race, intelligence, productivity, able-bodiedness, and capacity for enhancement continue to shape whose lives and futures are valued. Eugenics has not disappeared. It has been updated for an age in which powerful elites imagine redesigning humanity through technologies that reproduce existing hierarchies.

Techno-optimism turns this hierarchy into a story of liberation, presenting corporate technologies as neutral instruments of progress while concealing who owns them and who bears their costs. At its authoritarian edge, it overlaps with neo-reactionary or “Dark Enlightenment” thought, which rejects egalitarian democracy and imagines society governed like a corporation by CEOs or technocratic elites. Such an order need not announce itself as a dictatorship. It can emerge through privately owned systems that monitor, classify, restrict, and punish people without meaningful consent or appeal.

Silicon Valley’s supposed hostility to the federal government is selective. The industry resists the state as regulator while embracing it as investor, customer, military partner, police force, and border authority. It does not necessarily seek a smaller state, but one that protects a techno-capitalist oligarchy, suppresses resistance, and uses privately developed systems to govern.

That bargain has entered the machinery of the Trump administration. Reuters reported that technology executives and companies gave more than $300 million to support Trump’s 2024 campaign and an allied political committee. Elon Musk spent more than a quarter-billion dollars, while David Sacks and figures from Marc Andreessen’s network entered or advised the administration. Whether driven by ideology, profit, or access, this alliance advances domestic authoritarian priorities (including ideological policing, mass surveillance, and deportation) and an imperial project of military dominance and worldwide dependence on American technology. The state gains private infrastructure and expertise. The companies receive deregulation, subsidies, contracts, and support for global expansion.

The administration’s America’s AI Action Plan turns that bargain into policy through accelerated adoption, fewer regulatory barriers, rapid data center construction, and global export of American chips, models, software, and standards. A related executive order on “Preventing Woke AI” directs agencies to buy models conforming to the administration’s definitions of truth and neutrality while casting systemic racism, intersectionality, and what it calls “transgenderism” as ideological threats. The state is not removing politics from AI. It is using public purchasing power to present its own politics as neutral truth.

None of this means every signer or developer endorses the Trump administration or shares the ideological commitments described here. The point is structural: both letters operate within a political order built around private infrastructure, concentrated wealth, U.S. geopolitical dominance, and interstate competition for AI-enabled military and technological superiority. The race (especially between the United States and China) to develop AI for weapons, intelligence, surveillance, cyberwarfare, logistics, and military planning turns continued expansion into a national-security imperative and makes restraint appear to be strategic surrender. Open weights accelerate diffusion, frontier-safety governance can legitimize control by insiders, and public institutions supply the money, data, infrastructure, legitimacy, and users. AI spreads throughout society, but ownership, power, and wealth remain concentrated.



This article is licensed under Creative Commons (CC BY-NC-ND 4.0), and you are free to share and republish under the terms of the license.

Tim Scott is an Associate Professor of Social Work at Central Connecticut State University, a tech industry researcher, and the author of the book, Schooling for Silicon Valley.

Most Westerners know Buddhism through meditation, but everyday Buddhist life looks quite different in Thai temples

(The Conversation) — Thai Buddhism involves daily rituals and collective acts of worship, unlike the individual-focused emphasis on meditation often found in Western Buddhist communities.

Buddhist daily practice in Thailand involves many rituals. (Catherine Leblanc/Godong/Stone via Getty Images)

Brooke Schedneck

August 25, 2026 

(The Conversation) — I recently returned from leading an educational program in Thailand focusing on how Buddhists practice the religion in their day-to-day lives. The participants – Buddhist practitioners from North America and Europe – wanted to deepen their understanding and practice of the religion.

Many Western Buddhists’ practices centers on meditation. As a scholar of contemporary Thai Buddhism, I have taught many students and adults in the U.S. who have a similar understanding.

Although meditation can be very important to many practitioners, as I show in my 2023 book, “Living Theravada,” Buddhism is more than an intellectual tradition focused on the mind. In countries with a majority Buddhist population, such as Thailand, Buddhist practice involves rituals, offerings to monks and visits to temples, among others traditions.

Here are three ways in which Buddhism shapes daily life in Thailand.

1. Merit-making

A central part of daily practice is performing rituals associated with merit-making. To make merit means to do something morally good that will generate good karma – actions that are believed to impact this life and future ones. The simplest way to do this is to make a monetary donation to a temple.

There are often several merit-making activities in temples that are meant to be fun and engage participants. For example, one can place a small sheet of gold leaf on a part of a Buddha statue where one would like help or healing, such as one’s mouth if someone often uses their voice for their work or the ankle where they would like help or healing.

Buddhists believe that placing a sheet of gold leaf on the statue of a Buddha will help in healing.
Prapass Pulsub/Moment via Getty Images

Another fun activity is ritually purifying the “chedi,” or a large monument that enshrines sacred relics. Thai Buddhists fill a container with blessed water, which is attached to a pulley that leads to the upper part of the chedi. The container empties when it reaches the top, and participants make merit by symbolically cleansing the sacred monument.

Making offerings to specific Buddha statues representing the different zodiac signs based on one’s day and year of birth is common. In Thai temples, one can often find eight Buddha statues in a row, all in different poses; each represents a day of the week, with Wednesday being divided into morning and evening. Those born on a Friday, for example, will make an offering to the Buddha statue corresponding to their day of birth.

Worshipers can also make merit by directing donations to the zodiac animal associated with their birth year. Those born during the year of the monkey can place a flag or paper decorated with the monkey symbol on or around a sacred structure inside the temple hall. In other words, individuals can make merit connected to their own circumstances, relationships and aspirations, while practicing generosity.

2. Spiritual connections with monks

A foundational feature of Theravada Buddhism is the relationship between monastics and the lay people. This relationship is a delicate balance where laity offer alms while the monks offer spiritual wisdom and guidance. As monks’ lives are dedicated to Buddhism, they are believed to accrue much merit, which Buddhists believe they can then transfer to people who make offerings.

However, there are times when these bonds can be tested. During COVID-19, for example, the laity did not go into the temples for fear of contracting the virus. This made it difficult for monks to obtain food and maintain connections with their local Buddhist community.

The laity’s faith in monks and in their ability to provide merit and moral guidance can suffer if monks break their vows, such as by drinking alcohol or having sex. In 2022, for example, a well-known monk was found to have had an affair with a model. Thai Buddhists felt disappointed and raised questions about the ability of young monks to take their vow of celibacy seriously.

3. Rituals and wish-making

The ultimate goal of Buddhist practitioners is enlightenment, or nirvana – freedom from the cycle of birth and death and suffering. But most Buddhists see this goal as out of reach in this lifetime and instead have desires for worldly things, such as getting a job, winning the lottery or finding a spouse.

A statue of the Hindu god Ganesha at the entrance to the Wat Sri Suphan temple in Chiang Mai, Thailand.
© Marco Bottigelli/Moment via Getty Images

Thai Buddhist temples often have icons of Hindu gods, such as Ganesha and Brahma, who are believed to grant these worldly wishes. Some temples may also have shrines dedicated to indigenous gods or spirits of the land. Since these gods are believed to be concerned with worldly success, they are popular places where people pray for wealth, job promotion, romance or fertility.

These rituals involve an exchange, where worshipers promise to return with certain objects the god or spirit is believed to enjoy, such as hard-boiled eggs, figurines of elephants or flower garlands, if their wish comes true.

Buddhism in Thailand is embedded into everyday lives and collective concerns. At the end of 10 days, my program participants learned the value of merit-making and felt benefit from performing rituals. They saw how Buddhism extends beyond the individual pursuit of mental transformation through meditation.

(Brooke Schedneck, Associate Professor of Religious Studies, Rhodes College. The views expressed in this commentary do not necessarily reflect those of Religion News Service.)