The idea at the heart of the paradox is a familiar one in AI ethics, where researchers have long discussed value alignment, human compatibility, and social choice ethics as guiding frameworks. The democratic character of the approach comes from its structure: just as democracies use voting to aggregate citizen preferences into collective decisions, a democratic AI system aggregates the values of many people into a single measure that steers its behavior. The alternative is what Baum calls authoritarian or dictatorial AI, in which a narrow group — or even a single individual — imposes its values on the machine from the top down. On paper, the democratic model sounds like the obvious moral choice. But the paradox exposes a hidden trapdoor: if the aggregated values of the public turn against the AI itself, a truly democratic system has no principled way to keep running. Its own moral machinery becomes the instrument of its abolition.

Baum extends the logic from individual systems to whole research and development enterprises. Imagine a project that declares its mission to be the advancement of democracy, positioning its AI as a contribution to the global contest between democratic and authoritarian governance — yet the public, whose preferences the project claims to serve, would rather the project did not exist. By the project’s own democratic standard, it ought to disband. This is not a purely hypothetical scenario, Baum argues. The leadership of major AI laboratories, including OpenAI and Anthropic, both based in the United States, have publicly framed their work in explicitly democratic terms, casting advanced AI as a bulwark against authoritarian dictatorship. Meanwhile, American public opinion on AI is mixed and increasingly negative, and civic opposition is mounting against specific aspects of the industry, most visibly the proliferation of energy-hungry data centers. Baum is careful to note that this does not constitute an unambiguous case of unwanted democratic AI, but he argues that the signs point clearly in that direction.

Future scenarios could sharpen the tension dramatically. In a provocative essay titled “Eat your AI slop or China wins,” writer Ross Bellafiore argues that the United States should embrace aggressive AI development to outcompete China, even at the cost of severe harms — mass unemployment, retreat into delusional virtual worlds, and learned helplessness among people who can no longer function without tools like ChatGPT. If public opinion is already souring under comparatively mild conditions, Baum reasons, opposition would likely intensify sharply if conditions deteriorated along those lines. Yet the picture is not uniformly bleak, because AI is an enormously broad category. Surveys show that Americans strongly support AI applications in weather forecasting, drug development, and the detection of fraud and financial crime. It may follow, Baum suggests, that the public would favor narrow AI built for specific, beneficial purposes over the pursuit of artificial general intelligence, the hypothetical technology that dominates the ambitions of the largest laboratories.

That distinction matters because AGI is also heavily associated with catastrophic risk. Scenarios in which AI causes extreme catastrophe are, by definition, unlikely to win public support, and uncertainty about which technologies might trigger such outcomes forces democratic AI to confront hard questions about how to weigh catastrophic risks against other considerations. There is already evidence of where the public stands: strong bipartisan support has emerged for legislation addressing catastrophic AI risks, suggesting a low public appetite for existential gambles — and, by extension, for AGI itself rather than narrow applications. For any project that claims democratic legitimacy while racing toward the most powerful and least predictable forms of the technology, this gap between stated values and public preference represents the paradox in its most acute form.

The paradox also exposes how thin the industry’s conception of democracy really is. In political science, democracy understood simply as the aggregation of citizen preferences through elections is a minimalist, or “thin,” conception. Thicker conceptions add citizen participation in public deliberation and issue advocacy — and it is precisely these richer dimensions of democracy that AI may be undermining. Baum catalogs the mechanisms: AI can generate and spread misinformation at scale, including synthetic deepfake videos, degrading citizens’ ability to form accurate political preferences and hold governments accountable. AI-powered information platforms siphon advertising revenue from news media, further eroding that accountability. AI tools can generate and popularize ballot initiatives, crowding out traditional human-run advocacy organizations. And automation of the economy concentrates wealth and political power in the hands of the narrow elite that owns the technology. There are counterexamples — research has shown that AI chat interventions can make online political conversations less divisive — but Baum concludes that the literature suggests AI’s overall effect on democracy may be strongly negative. He reserves a pointed term for projects that profess democratic values while weakening democracy in practice: “democratic” AI, in scare quotes.

Complicating everything is geopolitics. Bellafiore’s essay articulates what Baum calls the greatest challenge for democratic AI: the prospect that AI becomes so powerful that an advantage in the technology could confer global domination, handing the future to whichever regime — democratic or authoritarian — gets there first. Under that assumption, even a “democratic” AI project that damages democracy at home might still be the lesser of two evils, the only hope of preserving any semblance of democratic governance against authoritarian rivals. But Baum finds this scenario deeply bleak, because it implies democracy is doomed to decline and the only question is how far the fall will go. Crucially, the scenario also hinges on contested assumptions. Current AI technology does not confer global domination, and the International AI Safety Report, led by Yoshua Bengio and colleagues, underscores how deeply uncertain and controversial the prospects for such power remain.

If future AI will not confer world domination, the grim binary dissolves. Competition over AI would then resemble competition in other economic sectors, such as renewable energy and electric vehicles, or in military technologies like drones and stealth aircraft — domains that matter, but that can be balanced against the health of democracy. If the choice is between a stronger democracy and market share in one industry, Baum argues it may be entirely reasonable to choose democracy by heavily regulating or even disbanding “democratic” AI projects. And even if AI could confer domination, he identifies a third path that industry leaders claim does not exist. Sam Altman and Bellafiore have both asserted there is no option between democratic and authoritarian AI, but Baum disagrees: the third option is diplomacy — multilateral or even global agreements to avoid AGI and other extreme AI technologies altogether. Diplomacy guarantees nothing, he concedes, but neither does the aggressive pursuit of advanced AI, and given AI’s potential to be unwanted and corrosive, diplomacy may be the only route to genuinely democratic outcomes.

Baum closes with the conflicts of interest that shadow the entire debate. AI laboratories are not neutral actors within democracies; they are private corporations pursuing market share and profit, and their institutional success can collide with the democratic values they espouse — as when they accept investments from authoritarian governments. Maintaining a pro-democracy image may help such companies dodge regulation and win government support, and if AI ever did become dominant, the incentive to drop the ruse would be overwhelming. From the citizen’s vantage point, Baum argues, this gives democracies strong reason to regulate or even ban AGI pursuits, or to nationalize them under full democratic control. His prescriptions are concrete: AI projects should restrain misinformation-enabling tools, fund independent journalism, support policies that reduce concentrations of wealth and power, back AI diplomacy, and be prepared to shut down if they are unwanted. Citizens, for their part, can organize into advocacy coalitions to push sound AI policy, demand corporate governance in the public interest, and fight the broader corruption of democracy by concentrated wealth. The answer to the paradox, in the end, is conditional: narrow, publicly supported AI can be democratic. Extreme AI pursued in democracy’s name may not deserve to exist at all.

Subject of Research: The paradox of democratic AI systems that must self-destruct if public aggregate values oppose their existence

Article Title: The paradox of unwanted democratic AI

Article References: Baum, S. D. (2026). The paradox of unwanted democratic AI. AI & SOCIETY. https://doi.org/10.1007/s00146-026-03386-y

Image Credits: AI Generated

DOI: 10.1007/s00146-026-03386-y

Keywords: artificial intelligence, AI ethics, democracy, value alignment, social choice, AGI, catastrophic risk, OpenAI, Anthropic, AI regulation, AI diplomacy, public opinion

The core argument is deceptively simple but far-reaching. AI is transforming how the social sciences observe, model, simulate, and govern human systems, from financial markets to classrooms. Yet at the same time, some of AI’s most stubborn problems—interpretability, controllability, legitimacy, and value alignment—are fundamentally social and institutional questions, not merely technical ones. Why did a model reach a decision? Who is accountable when an autonomous agent acts? What values should an algorithm optimize, and who decides? These are questions that social science has grappled with for centuries, and the authors contend that answering them requires social scientific theory to be built into AI systems from the ground up.

To formalize this two-way relationship, the paper introduces a framework the authors call AI × SS, which unifies two complementary research directions under a single system architecture. The first direction, AI4SS, uses artificial intelligence to advance social science research: machine learning to detect patterns in massive behavioral datasets, natural language processing to analyze text at scale, and simulation to test hypotheses about collective behavior. The second direction, SS4AI, runs the other way, importing insights from sociology, economics, psychology, political science, and law to make AI systems more transparent, governable, and aligned with human values. The multiplication sign in the framework’s name is deliberate: the authors see the combination as producing something qualitatively new rather than a simple sum of the two fields.

A central theoretical pillar of the framework is Parallel Intelligence, or PI, an approach associated with Fei-Yue Wang’s long-running research program on complex systems. In the PI paradigm, real social systems are paired with artificial, computational counterparts—virtual societies, synthetic populations, and simulated institutions—that run in parallel with reality. Data flows between the real and artificial systems, allowing researchers to conduct experiments that would be impossible or unethical in the real world, to predict the consequences of policy interventions before deploying them, and to prescribe better decisions by learning from the parallel worlds. Generative AI and foundation models have dramatically strengthened this vision, because large models can now generate plausible synthetic agents, scenarios, and social dynamics at unprecedented scale and fidelity.

The review organizes the current research landscape into four methodological pathways. The first is data representation: how human behavior, institutions, and social relations are encoded so that machines can process them, ranging from network representations of societies to multimodal embeddings of text, images, and interaction traces. The second is analysis and modeling, where techniques such as graph learning, causal inference, and large language model-based agents are used to extract structure and explanation from social data. The third is simulation and experimentation, encompassing agent-based modeling, generative agent societies, and the construction of computational laboratories for testing social theories. The fourth is decision and optimization, where models are turned into actionable guidance for governance, resource allocation, and policy design under uncertainty.

Along each of these pathways, the authors survey representative studies that illustrate how the frontier is moving. In data representation, the explosion of digital trace data—from social media, mobility records, and transaction logs—has created both opportunity and risk, demanding new methods for privacy preservation and bias correction. In analysis and modeling, foundation models pretrained on vast corpora are being repurposed as social simulators, capable of reproducing survey responses, opinion dynamics, and market behavior, though the authors caution that their fidelity and faithfulness remain open empirical questions. In simulation, generative agents that plan, remember, and interact are enabling synthetic societies in which researchers can stress-test interventions against thousands of virtual lives before touching the real world.

The paper then grounds these abstractions in concrete domains. In finance, AI-driven models now read markets, news flows, and investor sentiment in real time, while social science supplies the institutional understanding of regulation, trust, and systemic risk. In communication, recommendation algorithms and large language models are reshaping public discourse, and computational social science is racing to measure effects such as polarization and misinformation. In management, agentic AI systems are beginning to participate in organizational decision-making, raising questions about delegation, oversight, and human–machine collaboration that draw directly on organizational theory. In art, generative models are transforming creative production and raising questions of authorship and value. In education, adaptive AI tutors promise personalization at scale, but their legitimacy and fairness depend on pedagogical and ethical frameworks that only the social sciences can supply.

What emerges across all of these domains, the authors argue, is a new research structure rather than a loose collection of applications. Institutions, governance mechanisms, and human–machine collaboration are not afterthoughts bolted onto AI systems; they are constitutive components of any AI that operates in society. An autonomous agent acting in a market, a classroom, or a city is embedded in norms, incentives, and legal regimes, and its behavior cannot be predicted or controlled without modeling those social layers. Conversely, the arrival of agentic AI—systems that pursue goals, coordinate with other agents, and act over long horizons—creates genuinely novel social phenomena, such as economies of interacting algorithms, that demand new social scientific theory in their own right.

The review also reflects an unusually broad collaborative effort. The author team spans institutions including Macau University of Science and Technology, the Chinese Academy of Sciences, Beijing Jiaotong University, Tianjin University, the University of Science and Technology of China, Zhejiang Lab, Beijing Institute of Technology, the University of Chinese Academy of Sciences, and the University of Glasgow, and grew out of a series of workshops organized under the Artificial Intelligence for Autonomous Science initiative between 2023 and 2026, covering topics from social transportation to intelligent art and parallel decision theaters. That breadth mirrors the paper’s central claim: no single discipline owns the AI–society problem anymore.

For researchers and policymakers alike, the practical message is that the next generation of breakthroughs will come from teams that treat AI and the social sciences as a single coupled system. Building interpretable models requires theories of explanation that humans actually find meaningful. Building controllable agents requires institutional design, incentive analysis, and accountability structures. Building value-aligned systems requires explicit, contestable accounts of whose values are being encoded. The AI × SS framework offers a map of that territory, organizing data, modeling, simulation, and decision-making into a coherent architecture in which artificial and human systems learn from each other continuously. If the authors are right, the most important new social science of the coming decade may be the one that is co-written with machines.

Subject of Research: The bidirectional co-evolution of artificial intelligence and the social sciences through a unified AI × SS framework based on parallel intelligence

Article Title: Ai and social sciences: new AI, new studies, and new social sciences

Article References: Ni, Q., Lin, F., Zhang, H., Xue, X., Zhang, T., Lü, L., Zhang, B., Lu, Y., Li, D., Yang, L., Ye, P., Li, X., Guo, C., Huang, J., Zheng, X., Yu, H., & Wang, F.-Y. (2026). Ai and social sciences: new AI, new studies, and new social sciences. Artificial Intelligence Review. https://doi.org/10.1007/s10462-026-11693-5

Image Credits: AI Generated

DOI: 10.1007/s10462-026-11693-5

Keywords: artificial intelligence, social sciences, parallel intelligence, AI4SS, SS4AI, computational social science, agentic AI, generative AI, foundation models, value alignment, human-machine collaboration, AI governance