10 papers
Billions of Sketches Reveal Hidden Cultural Variation in Human Concepts
Arianna Pera, Mauro Martino, Nima Dehmamy +3
Claims about the universality of human concepts have been predominantly assessed through linguistic similarity across languages and cultures. However, words are effective as commun…
Failure of contextual invariance in large language models
Sagar Kumar, Ariel Flint, Luca Maria Aiello +1
Standard evaluation practices assume that large language model (LLM) outputs are stable when prompts are embedded in contextually equivalent discourses. Here, we test this assumpti…
How malicious AI swarms can threaten democracy: The fusion of agentic AI and LLMs marks a new frontier in information warfare
Daniel Thilo Schroeder, Meeyoung Cha, Andrea Baronchelli +19
Advances in AI offer the prospect of manipulating beliefs and behaviors on a population-wide level. Large language models and autonomous agents now let influence campaigns reach un…
Group size effects and collective misalignment in LLM multi-agent systems
Ariel Flint, Luca Maria Aiello, Romualdo Pastor-Satorras +1
Multi-agent systems of large language models (LLMs) are rapidly expanding across domains, introducing dynamics not captured by single-agent evaluations. Yet, existing work has most…
Ideology and polarization set the agenda on social media
Edoardo Loru, Alessandro Galeazzi, Anita Bonetti +6
The abundance of information on social media has reshaped public discussions, shifting attention to the mechanisms that drive online discourse. This study analyzes large-scale Twit…
Reply to "Emergent LLM behaviors are observationally equivalent to data leakage"
Ariel Flint Ashery, Luca Maria Aiello, Andrea Baronchelli
A potential concern when simulating populations of large language models (LLMs) is data contamination, i.e. the possibility that training data may shape outcomes in unintended ways…