Daniel Karell’s research while affiliated with Yale University and other places

How does online speech affect offline attacks? While a growing literature has examined this link in right-wing violence in the West, much less is known about its importance in the religiously divided societies of the Global South. Furthermore, existing research has overwhelmingly focused on negative externalities of social media, while paying comparatively little attention to their conciliatory effects. We advance the scholarship in both of these areas by analyzing 22.4 million posts from Koo, an Indian social media network popular among India's Hindu nationalists. We combine these data with information on attacks on religious minorities in India from 2020 through 2022. We find that the frequency of hashtags with a Hindu-chauvinist connotation are associated with increased attacks on Muslims and Christians. We also find that the frequency of hashtags alluding to the overcoming of religious divisions is associated with fewer attacks. These results survive a battery of robustness checks and supplemental tests. Additionally, the observed relationships disappear during exogenous Internet outages, consistent with the effect being driven by online speech. Importantly, since the content we study is not overtly aggressive and conveys values rather than factual claims, it does not classify as hate speech, misinformation, or disinformation. This suggests that the scholarly debate on what kinds of online speech influence offline harm has to be broadened and that censorship and fact-checking can fall short of addressing online speech's negative consequences in religiously divided societies.

Generative artificial intelligence (AI) offers new capabilities for analyzing data, creating synthetic media, and simulating realistic social interactions. This essay introduces a special issue that examines how these and other affordances of generative AI can advance social science research. We discuss three core themes that appear across the contributed articles: rigorous measurement and validation of AI-generated outputs, optimizing model performance and reproducibility via prompting, and novel uses of AI for the simulation of attitudes and behaviors. We highlight how generative AI enable new methodological innovations that complement and augment existing approaches. This essay and the special issue’s ten articles collectively provide a detailed roadmap for integrating generative AI into social science research in theoretically informed and methodologically rigorous ways. We conclude by reflecting on the implications of the ongoing advances in AI.

How do artificial intelligence (AI) tools affect people's knowledge? Existing scholarship supports contrasting expectations. Some research shows that people often find AI-generated ("synthetic") texts clearer, more accessible, and more enjoyable to read, which can foster learning from the texts. However, other research shows that synthetic texts are derivative and homogeneous, making them less interesting and thus potentially inimical to learning. In two experimental studies, we find that people who read AI-generated summaries of historical events more accurately recall factual information about the events compared to people who read human-written summaries of the same event. In the first study (N=193), participants who read synthetic text describing the Seattle General Strike of 1919 correctly answered more fact-based questions than participants who read a text written by experts. The preregistered second study, with a sample nationally representative of the United States (N=1907) and additionally testing a second historical event, replicates the results of the first study. Study 2 also indicates that synthetic historical summaries have a positive effect regardless of whether subjects know they were generated by AI, as well as whether or not the summaries exhibit political biases. Furthermore, while synthetic summaries improve knowledge for participants across all levels of educational attainment, the largest improvements occur at the lowest levels. Our findings show how AI-generated texts can enhance learning and reduce, rather than amplify, educational inequalities.

A large literature on international labor migration explores how to improve low-skilled migrants’ experience of pursuing and obtaining overseas employment. Much of this scholarship focuses on describing and mitigating difficult, and sometimes exploitative, conditions in the host country. Scholars have paid less attention to factors in home countries that may affect aspiring migrants’ experience of looking for overseas work. We help address this gap by conducting a field experiment in the high out-migration country of Pakistan to examine whether receiving new information about employment brokers and overseas opportunities improves aspiring migrants’ subjective and objective experience of the job-seeking process. After analyzing the effects of information on those who lack alternative sources of information, we report mixed findings. These highlight the need to think carefully about ways to improve aspiring migrants’ job-seeking experiences, especially with regard to the common assumption that if low-skilled migrants simply “knew more” — that is, had more relevant and accurate information — they would not allow themselves to have negative migration experiences; our findings suggest this is not necessarily always the case.

Leaders are a part of virtually every group and organization, and while they help solve the various collective action problems that groups face, they can also be unprincipled and incompetent, pursuing their own interests over those of the group. What types of circumstances foster prosocial leadership and motivate leaders to pursue group interests? In a modified dictator game (N = 798), we examine the effects of piece-rate subsidies (or pay per unit of work performed) and the relative price of giving (or the size of the benefit to others for giving) on prosocial behaviour and norms about giving. We find that subsidies increase giving by leaders, that the relative price of giving is unrelated to prosocial behaviour, and that neither affects norms about giving. Furthermore, the introduction and removal of a subsidy do not undermine giving over time. Our results imply that subsidies increase group welfare by motivating leaders to allocate a larger share of resources to group members.

How does social media content affect users’ online discourse? Existing scholarship sheds light on how several social media features involving content can influence users’ speech. However, this research often conflates content’s lexical dimension with its symbolic dimension. The authors analyze how the symbolic properties of online content can distinctly affect discourse on social media. Specifically, they examine how the symbolic meanings conveyed by Twitter’s reinstatement of Donald Trump’s account influenced Twitter users’ discourse. The results of embedding regressions indicate that Trump’s reinstatement immediately shifted users’ discourse about social and political identity-based groups, but only when they discussed Black and Jewish people. Additional results suggest that the discourse became more politicized and that the discursive shift was short lived. The authors’ findings contribute conceptual and analytical clarity to the sociosemantic dynamics of online discourse, encouraging future research to distinguish and compare the lexical and symbolic dimensions of online content.

The development of generative artificial intelligence (genAI) has caused concern about its potential risks, including how its ability to generate human-like texts could affect our shared perception of the social world. Yet, it remains unclear how best to assess and understand genAI's influence on our understanding of social reality. Building on insights into the representation of social worlds within texts, we develop an initial framework for analyzing genAI's content and its consequences for perceptions of social reality. We demonstrate this "synthetic duality" framework in two parts. First, we show that genAI can create, with minimal guidance, reasonable portrayals of actors and ascribe relational meaning to those actors – virtual social worlds within texts, or "Mondo-Breigers". Second, we examine how these synthetic documents with interior social worlds affect readers’ view of social reality. We find that they change individuals' perceptions of actors depicted in the documents, likely by updating individuals' expectations about the actors and their meanings. However, additional exploratory analyses suggest it is texts' style, not their construction of "Mondo-Breigers", that might be influencing people's perceptions. We end with a discussion of theoretical and methodological implications, including how genAI may unsettle structural notions of individuality. Namely, reimagining the duality of individuals and groups could help theorize growing homogeneity in an increasingly AI-informed world.

Early modern economic thought is often left out of histories of economic thought, but it was an important inflection point in the development of the modern science of economics. Using network clustering and topic modeling, this research describes the subjects covered in the economic literature of the early modern era and how they change over time using a sample of 2,353 economics texts written from 1550 to 1700s. Topic modeling of the corpus produces sixty-six distinct topics. These topics are clustered using network science methods that reveal five main themes: trade, politics, travel, religion, and husbandry. Longitudinal analysis of the prevalence of topics shows increases in politics and decreases in travel, religion, and husbandry. The results also allow a more precise sense of when those shifts take place than has previously been established, implying that computational methods hold great promise for revealing new information and insights about the evolution and transformation of economic thought over time.