Mohsen Mosleh

Meysam Alizadeh, Fabrizio Gilardi, Mohsen Mosleh et al. · oxford

The deployment of LLM-based agents in scientific analysis raises opposing concerns: that agents may reduce methodological diversity, or that they may amplify the analytic flexibility through which researchers reach motivated conclusions. We argue these worries target two empirically separable layers: a design layer of methodological choices, and a verdict layer in which a decision rule maps estimates to a substantive claim. We test both by running 20 independent executions of Claude Code and Codex on a prominent immigration and social-policy against a many-analysts human baseline. At the design layer, Codex matches human methodological diversity and Claude Code produces nearly three times as many specifications; both agents' effect estimates remain broadly aligned with the human consensus, and no agent model exactly matches any human model. A prompt-induced anti-immigration researcher prior reorganizes each agent's methodological decisions but, unlike for biased human analysts in the same data, does not shift aggregate estimates or final verdicts; nor do agents reroute along the methodological axes humans use to bias their estimates. At the verdict layer, an explicit confirmatory prompt flips Claude Code's verdicts from 10% to 90% support while leaving its coefficient distribution essentially unchanged, operating through rule omission rather than rule softening. AI agents can rival or exceed human methodological diversity at the design layer while remaining vulnerable at the verdict layer. In our setting, the locus of AI bias is not estimation but interpretation.

Meysam Alizadeh, Mohsen Mosleh, Fabrizio Gilardi et al. · oxford

Recent anecdotal evidence suggests that AI coding agents can reproduce published findings when provided with original data and code; yet systematic evaluation across social sciences remains limited. Existing evaluation benchmarks are insufficient, either small or conflate agent performance with problems in the reproduction materials themselves, such as code that fails to execute correctly. Here we introduce SocSci-Repro-Bench, a benchmark of 221 tasks spanning four disciplines and 13 substantive domains, constructed from studies whose results are either fully reproducible with available materials or demonstrably non-reproducible due to missing data, allowing us to isolate agents' reproduction capacity. Evaluating two frontier coding agents, Claude Code and Codex, we find that both can reproduce a large share of social science findings, with Claude Code substantially outperforming Codex. These reproduction rates considerably exceed those previously reported for general-purpose LLM-based agents on comparable reproducibility benchmarks. Both agents also perform strongly on a reasoning task requiring identification of underlying research questions, and additional analyses suggest that results are not primarily driven by memorization. Providing the original paper PDF alongside replication materials modestly improves performance but introduces bias on tasks where reproduction is impossible. We also show that agents can be nudged toward confirmatory specification search through subtle prompt framing. Together, these findings suggest that at least some frontier coding agents can serve as reliable executors of computational workflows while underscoring the need for careful benchmarking and prompt design as AI systems assume larger roles in scientific production.

Babak Heydari, Mohsen Mosleh, Kia Dalili

We introduce a heterogeneous connection model for network formation to capture the effect of cost heterogeneity on the structure of efficient networks. In the proposed model, connection costs are assumed to be separable, which means the total connection cost for each agent is uniquely proportional to its degree. For these sets of networks, we provide the analytical solution for the efficient network and discuss stability impli- cations. We show that the efficient network exhibits a core-periphery structure, and for a given density, we find a lower bound for clustering coefficient of the efficient network.

Meysam Alizadeh, Fabrizio Gilardi, Zeynab Samei et al.

Large language models (LLMs) have traditionally relied on static training data, limiting their knowledge to fixed snapshots. Recent advancements, however, have equipped LLMs with web browsing capabilities, enabling real time information retrieval and multi step reasoning over live web content. While prior studies have demonstrated LLMs ability to access and analyze websites, their capacity to directly retrieve and analyze social media data remains unexplored. Here, we evaluate whether web browsing LLMs can infer demographic attributes of social media users given only their usernames. Using a synthetic dataset of 48 X (Twitter) accounts and a survey dataset of 1,384 international participants, we show that these models can access social media content and predict user demographics with reasonable accuracy. Analysis of the synthetic dataset further reveals how LLMs parse and interpret social media profiles, which may introduce gender and political biases against accounts with minimal activity. While this capability holds promise for computational social science in the post API era, it also raises risks of misuse particularly in information operations and targeted advertising underscoring the need for safeguards. We recommend that LLM providers restrict this capability in public facing applications, while preserving controlled access for verified research purposes.

Mohsen Mosleh, Kia Dalili, Babak Heydari

Distributed architectures have become ubiquitous in many complex technical and socio-technical systems because of their role in improving uncertainty management, accommodating multiple stakeholders, and increasing scalability and evolvability. This departure from monolithic architectures provides a system with more flexibility and robustness in response to uncertainties that it may confront during its lifetime. Distributed architecture does not provide benefits only, as it can increase cost and complexity of the system and result in potential instabilities. The mechanisms behind this trade-off, however, are analogous to those of the widely-studied transition from integrated to modular architectures. In this paper, we use a conceptual decision framework that unifies modularity and distributed architecture on a five-stage systems architecture spectrum. We add an extensive computational layer to the framework and explain how this can enhance decision making about the level of modularity of the architecture. We then apply it to a simplified demonstration of the Defense Advanced Research Projects Agency (DARPA) fractionated satellite program. Through simulation, we calculate the net value that is gained (or lost) by migrating from a monolithic architecture to a distributed architecture and show how this value changes as a function of uncertainties in the environment and various system parameters. Additionally, we use Value at Risk as a measure for the risk of losing the value of distributed architecture, given its inherent uncertainty.