Academic Research

Millions of people around the world query large language models (LLMs) for information. Although several studies have compellingly documented the persuasive potential of these models, there is limited evidence of who or what influences the models themselves, leading to a flurry of concerns about which companies and governments build and regulate the models. Here we show through six studies that government control of the media across the world already influences the output of LLMs via their training data. We use a cross-national audit to show that LLMs exhibit a stronger pro-government valence when prompted in the languages of countries with lower media freedom than in those with higher media freedom. This result is correlational, so to triangulate the specific mechanism of how state media control can influence LLMs, we develop a multi-part case study on China’s media. We demonstrate that media scripted and curated by the Chinese state appears in LLM training datasets. To evaluate the plausible effect of this inclusion, we use an open-weight model to show that additional pretraining on Chinese state-coordinated media generates more positive answers to prompts about Chinese political institutions and leaders. We link this phenomenon to commercial models through two audit studies demonstrating that prompting models in Chinese generates more positive responses about China’s institutions and leaders than do the same queries in English. The combination of influence and persuasive potential across languages suggests the troubling conclusion that states and powerful institutions have increased strategic incentives to leverage media control in the hopes of shaping LLM output.

This forthcoming edited volume (Cambridge University Press) examines the transformative impact of artificial intelligence on democratic institutions, political behavior, governance, and the discipline of political science itself. The volume represents the report of the American Political Science Association’s Presidential Task Force on AI, Politics, and Political Science, co-chaired by Joshua Tucker and Nathaniel Persily. 

Across twelve chapters produced by close to 60 scholars, the report evaluates how generative AI and machine learning systems are reshaping public opinion formation, political communication, labor markets, electoral processes, state capacity, and regulatory frameworks. The authors analyze both the opportunities and risks posed by AI technologies, including concerns surrounding information integrity, ideological personalization, surveillance, democratic accountability, and concentrated technological power. Themes that cut across multiple chapters include: the unprecedented power of a small number of AI corporations; the opacity and non-replicability of model outputs; bias in AI systems; and the absence of agreed-upon benchmarks for evaluation.The volume also addresses methodological and ethical implications for political science research, emphasizing transparency, reproducibility, and the responsible integration of AI tools into scholarly inquiry. Ultimately, the volume argues that AI will not only alter political institutions and citizen-state relations, but also may fundamentally reshape how political knowledge is produced and interpreted. It calls for sustained interdisciplinary collaboration and evidence-based governance to ensure that AI development supports democratic resilience rather than undermining it.

Large Language Models (LLMs) are increasingly deployed to curate and rank human-created content, yet the nature and structure of their biases in these tasks remains poorly understood: which biases are robust across providers and platforms, and which can be mitigated through prompt design. We present a controlled simulation study mapping content selection biases across three major LLM providers (OpenAI, Anthropic, Google) on real social media datasets from Twitter/X, Bluesky, and Reddit, using six prompting strategies (\textit{general}, \textit{popular}, \textit{engaging}, \textit{informative}, \textit{controversial}, \textit{neutral}). Through 540,000 simulated top-10 selections from pools of 100 posts across 54 experimental conditions, we find that biases differ substantially in how structural and how prompt-sensitive they are. Polarization is amplified across all configurations, toxicity handling shows a strong inversion between engagement- and information-focused prompts, and sentiment biases are predominantly negative. Provider comparisons reveal distinct trade-offs: GPT-4o Mini shows the most consistent behavior across prompts; Claude and Gemini exhibit high adaptivity in toxicity handling; Gemini shows the strongest negative sentiment preference. On Twitter/X, where author demographics can be inferred from profile bios, political leaning bias is the clearest demographic signal: left-leaning authors are systematically over-represented despite right-leaning authors forming the pool plurality in the dataset, and this pattern largely persists across prompts.

Generative artificial intelligence agents are becoming increasingly active participants in conversations on social media platforms, yet little is known about how they shape public discussion of social problems. We present two preregistered online experiments testing AI-generated summaries in simulated, interactive social media environments. AI summaries increased the quality of user comments, without systematically increasing toxicity or negative affect. At the same time, AI exposure reduced engagement with conversation threads. AI summaries also increased the semantic similarity between user comments and the AI-generated summaries, suggesting that these systems function as informational anchors that shape discussion. Together, the findings reveal a tradeoff: AI-generated summaries can improve conversation quality while narrowing conversational engagement and channeling how users articulate political arguments. These results speak to growing concerns about how embedded AI systems fundamentally alter platform dynamics and shape public discourse.

Large language models (LLMs) are increasingly used as synthetic survey respondents, yet it is unclear whether their belief-system structure matches that of real publics. We compare 28 LLMs to the 2024 General Social Survey (GSS) using 52 attitude items and demographic persona traits. We estimate polychoric correlation matrices and propagate un-certainty in the GSS via bootstrap resampling with multiple imputation. Constraint is measured by the variance share explained by the first principal component and by effective dependence, a determinant-based measure of global linear dependence. Across models, LLM personas exhibit substantially higher constraint than humans; conditioning on persona traits reduces constraint far more for LLMs, indicating greater demographic mediation. Projection onto a shared GSS basis further shows overemphasis of the leading dimension and missing secondary structure. These results caution against treating LLM personas as a reliable foundation for synthetic survey data generation.

Existing text scoring methods require a large corpus, struggle with short texts, or require hand-labeled data. We develop a text scoring framework that leverages generative large language models (LLMs) to (1) set texts against the backdrop of information from the near-totality of the web and digitized media, and (2) effectively transform pairwise text comparisons from a reasoning problem to a pattern recognition task. Our approach, concept-guided chain-of-thought (CGCoT), utilizes a chain of researcher-designed prompts with an LLM to generate a concept-specific breakdown for each text, akin to guidance provided to human coders. We then pairwise compare breakdowns using an LLM and aggregate answers into a score using a probability model. We apply this approach to better understand speech reflecting aversion to specific political parties on Twitter, a topic that has commanded increasing interest because of its potential contributions to democratic backsliding. We achieve stronger correlations with human judgments than widely used unsupervised text scoring methods like Wordfish. In a supervised setting, besides a small pilot dataset to develop CGCoT prompts, our measures require no additional hand-labeled data and produce predictions on par with RoBERTa-Large fine-tuned on thousands of hand-labeled tweets. This project showcases the potential of combining human expertise and LLMs for scoring tasks.

Existing approaches to estimating politicians' latent positions along specific dimensions often fail when relevant data is limited. We leverage the embedded knowledge in generative large language models (LLMs) to address this challenge and measure lawmakers' positions along specific political or policy dimensions. We prompt an instruction/dialogue-tuned LLM to pairwise compare lawmakers and then scale the resulting graph using the Bradley-Terry model. We estimate novel measures of U.S. senators' positions on liberal-conservative ideology, gun control, and abortion. Our liberal-conservative scale, used to validate LLM-driven scaling, strongly correlates with existing measures and offsets interpretive gaps, suggesting LLMs synthesize relevant data from internet and digitized media rather than memorizing existing measures. Our gun control and abortion measures -- the first of their kind -- differ from the liberal-conservative scale in face-valid ways and predict interest group ratings and legislator votes better than ideology alone. Our findings suggest LLMs hold promise for solving complex social science measurement problems.