arXiv subjects

Philine Widmer

Publications and source records attributed to Philine Widmer.

7 recordsLinked to original sources

arXiv2026

Who Owns the Online Media?

Ownership matters for the media's watchdog role. We map the ownership networks behind thousands of online news outlets in the U.S., Canada, and Europe. The networks reveal who is ultimately responsible for the news: for over half of the outlets, a single entity. The rest sit behind multi-layered structures, making responsibility hard to trace. Market concentration, measured comparably across countries, is largely low to moderate. Looking at content, we find that co-owned outlets report more similarly, even within fixed outlet pairs, as ownership changes -- not least in the U.S., where reader demand is often thought dominant.

econ.GN↗

arXiv2026

Science under Threat? A Natural Experiment in Economics

Academic freedom has come under growing strain worldwide. To study whether and how academics respond to political pressure, we exploit a natural experiment: the publication in early 2025 of a "blacklist" of words flagged by the U.S. government. We find that the release of this list led to a sharp reduction in the use of these flagged words among economists at universities that rely heavily on federal funding, relative to scholars from institutions that are less dependent on federal funding or based in the UK. The drop is driven by content related to gender, race, and environment. We show that changes are not simply semantic but reflect actual paper content and that neither the individual funding status nor time-invariant author characteristics are driving the effects. We also document interesting heterogeneous effects by department quality and author gender and ethnicity. Our findings are consistent with the idea that scholars respond strongly to political pressure.

econ.GN↗

arXiv2025

Gen-Review: A Large-scale Dataset of AI-Generated (and Human-written) Peer Reviews

How does the progressive embracement of Large Language Models (LLMs) affect scientific peer reviewing? This multifaceted question is fundamental to the effectiveness -- as well as to the integrity -- of the scientific process. Recent evidence suggests that LLMs may have already been tacitly used in peer reviewing, e.g., at the 2024 International Conference of Learning Representations (ICLR). Furthermore, some efforts have been undertaken in an attempt to explicitly integrate LLMs in peer reviewing by various editorial boards (including that of ICLR'25). To fully understand the utility and the implications of LLMs' deployment for scientific reviewing, a comprehensive relevant dataset is strongly desirable. Despite some previous research on this topic, such dataset has been lacking so far. We fill in this gap by presenting GenReview, the hitherto largest dataset containing LLM-written reviews. Our dataset includes 81K reviews generated for all submissions to the 2018--2025 editions of the ICLR by providing the LLM with three independent prompts: a negative, a positive, and a neutral one. GenReview is also linked to the respective papers and their original reviews, thereby enabling a broad range of investigations. To illustrate the value of GenReview, we explore a sample of intriguing research questions, namely: if LLMs exhibit bias in reviewing (they do); if LLM-written reviews can be automatically detected (so far, they can); if LLMs can rigorously follow reviewing instructions (not always) and whether LLM-provided ratings align with decisions on paper acceptance or rejection (holds true only for accepted papers). GenReview can be accessed at the following link: https://anonymous.4open.science/r/gen_review.

cs.LG↗

arXiv2025

The Dark Side of the Web: Towards Understanding Various Data Sources in Cyber Threat Intelligence

Cyber threats have become increasingly prevalent and sophisticated. Prior work has extracted actionable cyber threat intelligence (CTI), such as indicators of compromise, tactics, techniques, and procedures (TTPs), or threat feeds from various sources: open source data (e.g., social networks), internal intelligence (e.g., log data), and ``first-hand'' communications from cybercriminals (e.g., underground forums, chats, darknet websites). However, "first-hand" data sources remain underutilized because it is difficult to access or scrape their data. In this work, we analyze (i) 6.6 million posts, (ii) 3.4 million messages, and (iii) 120,000 darknet websites. We combine NLP tools to address several challenges in analyzing such data. First, even on dedicated platforms, only some content is CTI-relevant, requiring effective filtering. Second, "first-hand" data can be CTI-relevant from a technical or strategic viewpoint. We demonstrate how to organize content along this distinction. Third, we describe the topics discussed and how "first-hand" data sources differ from each other. According to our filtering, 20% of our sample is CTI-relevant. Most of the CTI-relevant data focuses on strategic rather than technical discussions. Credit card-related crime is the most prevalent topic on darknet websites. On underground forums and chat channels, account and subscription selling is discussed most. Topic diversity is higher on underground forums and chat channels than on darknet websites. Our analyses suggest that different platforms may be used for activities with varying complexity and risks for criminals.

cs.CR↗

arXiv2025

Media Slant is Contagious

This paper examines the diffusion of media slant. We document the influence of Fox News Channel (FNC) on the partisan slant of local newspapers in the U.S. over the years 1995-2008. We measure the political slant of local newspapers by scaling the news article texts to Republicans' and Democrats' speeches in Congress. Using channel positioning as an instrument for viewership, we find that higher FNC viewership causes local newspapers to adopt more right-wing slant. The effect emerges gradually, only several years after FNC's introduction, mirroring the channel's growing influence on voting behavior. A main driver of the shift in newspaper slant appears to be a change in local political preferences.

econ.GN↗

arXiv2024

Aligning Large Language Models with Diverse Political Viewpoints

Large language models such as ChatGPT exhibit striking political biases. If users query them about political information, they often take a normative stance. To overcome this, we align LLMs with diverse political viewpoints from 100,000 comments written by candidates running for national parliament in Switzerland. Models aligned with this data can generate more accurate political viewpoints from Swiss parties, compared to commercial models such as ChatGPT. We also propose a procedure to generate balanced overviews summarizing multiple viewpoints using such models. The replication package contains all code and data.

cs.CL↗

arXiv2022

RELATIO: Text Semantics Capture Political and Economic Narratives

Social scientists have become increasingly interested in how narratives -- the stories in fiction, politics, and life -- shape beliefs, behavior, and government policies. This paper provides an unsupervised method to quantify latent narrative structures in text documents. Our new software package RELATIO identifies coherent entity groups and maps explicit relations between them in the text. We provide an application to the United States Congressional Record to analyze political and economic narratives in recent decades. Our analysis highlights the dynamics, sentiment, polarization, and interconnectedness of narratives in political discourse.

econ.GN↗