Many people now use AI chatbots to obtain summaries of complex topics, yet we know little about how this affects knowledge acquisition, including how the effects might vary across different groups of people. We conducted two experiments comparing how well people recalled factual information after reading AI-generated or human-written historical summaries. Participants who read AI-generated summaries scored significantly higher on knowledge tests than those who read expert-written blog posts (Study 1) or Wikipedia articles (Study 2). These improvements were present regardless of whether readers knew the content was AI-generated or if the AI summaries were politically biased. Moreover, AI summaries improved recall across various demographic groups, including gender, race, income, education, and digital literacy levels. This suggets that using AI tools for everyday factual queries does not create new knowledge inequalities but could still amplify existing ones through differential access. Our findings indicate that the increasingly routine use of AI for information-seeking could enhance factual learning, with implications for education policy and addressing inequality.

This study attempts to advance automated content analysis from consensus-oriented to coordination-oriented practices, thereby embracing diverse coding outputs and exploring the dynamics among differential perspectives. As an exploratory investigation, we evaluate six GPT-4o configurations to analyze sentiment in Fox News and MSNBC transcripts on Biden and Trump during the 2020 U.S. presidential campaign. By assessing each model’s alignment with partisan perspectives, we explore how partisan selective processing can be identified in LLM-Assisted Content Analysis (LACA). The findings indicate that LLM-based partisan perspective simulations reflect politically polarized standpoints across partisan groups, revealing a pronounced divergence in sentiment analysis between Democrat-aligned and Republican-aligned persona models. This pattern is evident in intercoder-reliability metrics, which are higher among same-partisan than cross-partisan persona model pairs. Results also suggest that LLM partisan simulations exhibit stronger ideological biases when analyzing politically congruent content. This approach enhances the nuanced understanding of LLM outputs and advances the integrity of AI-driven social science research and may also enable simulations of real-world implications.

Systematic reviews are essential for evidence synthesis but often require extensive time and resources, especially during data extraction. This proof-of-concept study evaluates the performance of Elicit , an AI tool specifically developed to support systematic reviews, in the context of a systematic review on psychological factors in dermatological conditions. We compared Elicit’s automated data extraction with manually extracted data across 43 studies and 602 data points. Both were assessed against a consensus-based ground truth. Elicit achieved an overall accuracy of 81.4%, compared to 86.7% for human reviewers—a difference that was not statistically significant. In cases where Elicit and the human reviewer extracted the same information, this information was correct in 100% of instances, suggesting that agreement between human and machine may serve as a reliable proxy for validity. Based on these results, we propose a semi-automated workflow in which Elicit functions as a second reviewer, reducing workload while maintaining high data quality. Our results demonstrate that domain-specific AI tools can effectively augment data extraction in systematic reviews, especially in settings with limited time or personnel.

Research shows that concurrent and sequential self-administered mixed-mode designs both have advantages and disadvantages in terms of panel survey recruitment and maintenance. Since concurrent mixed-mode designs usually achieve higher initial response rates at lower bias than sequential mixed-mode designs, the former may be ideal for panel recruitment. However, concurrent designs produced high share of paper respondents relative to web respondents. Since these paper respondents have been found to be at higher risk of attrition, cause higher data collection costs, and slow down the fieldwork process, sequential mixed-mode designs may be more practical in the regular course of the panel study after recruitment. Our study provides experimental evidence on the effect of switching a panel study from concurrent to sequential mixed-mode design after the panel recruitment. Results show that this switch significantly increases the share of online respondents without harming response rates. Respondents who are pushed to the web by the design change differ significantly from respondents who continue to participate via paper questionnaires with regard to a number of socio-digital inequality correlates. This suggests that, while the share of online respondents can be increased through mode sequencing, keeping the paper mail mode option is vital for ensuring continued representation of societal subgroups.

Data breaches are an everyday occurrence, exposing the personal details of millions globally. The victim impacts of data breaches can be considerable, including a range of financial harms such as fraud and identity crime, as well as non-financial harms, such as declines in emotional and psychological wellbeing. While these harms are documented, there is less research exploring how data breaches in particular expose victims to further victimisation, specifically through phishing attacks by offenders. Using survey data from 2,019 victims of the Optus and Medibank/AHM data breaches in Australia in 2022, this article examines factors which relate to phishing attempts on these individuals. Results indicate limited factors in predicting those targeted by phishing attempts. This highlights the opportunistic nature of phishing attacks in the aftermath of these two data breaches and a more generalised approach taken by offenders to gain additional details. It also demonstrates a need for continued community education and awareness to protect further personal information from being accessed by offenders into the future.

Data donation, a research approach in which users voluntarily contribute their personal digital data, offers a solution to the limitations of traditional self-reporting and digital trace methods by enabling the collection of comprehensive, ethically sourced usage information across multiple devices and digital interfaces. However, this promising method remains underutilized due to low participation rates. Therefore, this review pursued two integrated aims. First, to synthesize evidence on factors that influence three forms of participation: hypothetical willingness (stated intention in imagined scenarios), actual willingness (consent to donate when asked), and successful completion (following through with the full donation process). Second, to appraise existing workflows, frameworks, and methodological tools and integrate this appraisal with the factor synthesis to derive best practices for improving participation. We synthesized 35 articles, of which 14 examined factors influencing participation and 21 provide methodological guidance. Five key factors were identified: sensitivity of the data, privacy concerns, perceived autonomy and control over the donation process, complexity of the process, and participant characteristics. To overcome barriers related to these factors, we recommend maximizing participant privacy through robust data donation frameworks, enhancing transparency and user-friendliness, empowering participants by increasing autonomy and control over their data, and proactively addressing potential selection biases.

The news media are a primary but limited source of political information for voters. This article examines how the media coverage of members of parliament (MPs) prompts people to seek more information about MPs—an important element of voters’ political knowledge and democratic accountability. I use all available online content from four major Estonian newspapers during 2015–2023 ( ∼ 140,000 articles) and pre-trained transformer models to classify the sentiment and policy issues of news articles that mention MPs. I match media data with MPs’ Wikipedia pageviews on a daily basis over two legislative terms. First, the results show that media visibility of MPs is associated with more views of their Wikipedia page on the same day. Second, news articles with negative sentiment have a greater impact on information-seeking than those with a positive or neutral tone. Third, the impact of MPs' media appearances on information-seeking is dependent on the policy context as well as party affiliation. The findings have implications for understanding the consequences of political communication and democratic representation in the digital age.

We examine incivility across twenty high-traffic political and news subreddits to test how platform governance and social identity cues relate to on-platform discourse and participation. Building on theories of democratic communication and incivility, platform affordances and moderation, and uses-and-gratifications/network externalities, we specify how decentralized community rules and explicit in-group orientations could shape both the prevalence of uncivil language and patterns of engagement. We analyze a year-long, random sample of submissions and comments scored with established computational measures of incivility, and we link these scores to subreddit-level rule regimes and identity signaling. By distinguishing interpersonal impoliteness from democratic norm violations and by evaluating moderation complexity at the community level, this work clarifies when and how community governance relates to discourse quality and participation dynamics on Reddit. Findings inform ongoing debates about the efficacy of hybrid, human-centered moderation and the role of explicit identity norms in large online communities.

Factorial survey experiments are widely used in the social sciences to study decision-making and attitudes through controlled, experimentally manipulated scenarios – typically presented as text. While textual vignettes offer flexibility and ease of use, they often lack realism and may limit participant engagement. This article explores how generative artificial intelligence (AI) can be used to create customisable images for visual vignettes. It demonstrates techniques for producing and selectively editing images, highlighting their potential to address the demands of experimental social science research, while it also acknowledges key challenges, including ethical considerations, biases inherent in AI tools, and technical limitations. The article showcases potential applications of AI-generated images in social science research and draws on a pretest with human participants to present evidence on how AI-generated images are perceived and interpreted. By critically evaluating both opportunities and challenges, this article provides researchers with practical guidance on integrating AI-generated visuals into factorial survey experiments, enhancing methodological approaches in the social sciences.

With generative artificial intelligence’s (GenAI) growing popularity, individuals are increasingly using it when searching for election-related information. This scenario raises concerns that GenAI usage may result in widespread dissemination of misinformation, given its ability to generate seemingly authentic information. Nevertheless, despite the importance of Gen AI, few researchers have examined how individuals use this tool to search for election-related information. This study aims to assess how GenAI’s perceived system (i.e., accessibility and integration) and information quality (i.e., completeness, accuracy, and neutrality) impact its usage. Focusing on the 2024 U.S. presidential election, we conducted a two-wave survey and data was collected from 364 Americans. Participants were found to have a favorable attitude overall toward GenAI. Further, accuracy and neutrality were positively associated with GenAI usage. A fuzzy set qualitative comparative analysis was also conducted to identify different configurations of perceived system and information quality that led to high GenAI usage. Analyzing the qualitative responses further confirmed the results. This study contributes to the literature on the role of GenAI during elections, providing a nuanced understanding of how dimensions of GenAI’s perceived system and information quality impact individuals’ GenAI usage. The findings have significant practical implications for dealing with the (mis)information generated by GenAI.