Amplification to Synthesis: A Comparative Analysis of Cognitive Operations Before and After Generative AI

Problem
This preprint addresses the evolving landscape of cognitive operations in the geopolitical sphere, particularly in the context of the 2016 and 2024 U.S. presidential elections. While prior research has focused on bot-driven amplification of narratives, the advent of generative AI introduces new capabilities that may significantly alter the design and execution of these operations. The authors highlight a gap in understanding how generative AI influences cognitive operations, emphasizing the need for empirical analysis to inform mitigation strategies against potential threats posed by these technologies.

Method
The authors conducted a comparative analysis of behavioral and linguistic coordination patterns in X (formerly Twitter) datasets from the 2016 and 2024 elections, utilizing a combined corpus of over 133,000 posts. The methodology included several analytical techniques: post-type distribution analysis to categorize content types, semantic clustering to identify thematic groupings, temporal synchrony analysis to assess timing of posts, and Jaccard-based lexical overlap measures to evaluate the similarity of language used. These methods allowed for a comprehensive examination of shifts in content generation and narrative strategies between the two election cycles.

Results
The analysis revealed significant changes in the nature of posts between the two election years. Original content increased dramatically from 59% in 2016 to 93% in 2024, indicating a shift away from retweeting behavior. The mean Jaccard score for lexical overlap dropped from 0.99 to 0.27, suggesting that while posts were focused on similar topics, they employed markedly different language. Additionally, temporal coordination shifted from a model characterized by cross-semantic synchrony to one where posts co-occurred narratively, indicating a more concentrated and targeted approach to content dissemination. These findings suggest a transition towards operational strategies that leverage active content generation and narrative-specific targeting, likely influenced by generative AI technologies.

Limitations
The authors acknowledge that their analysis is limited to a specific social media platform and election context, which may not generalize to other platforms or geopolitical scenarios. They also note the potential for confounding variables that could influence the observed patterns, such as changes in user behavior or platform algorithms. Furthermore, the study does not explore the motivations behind the content generation or the actors involved, which could provide additional insights into the operational logic of cognitive operations.

Why it matters
This research has significant implications for understanding the role of generative AI in shaping cognitive operations and public discourse. By establishing an empirical baseline for the differences in content generation and narrative strategies before and after the introduction of generative AI, the findings can inform future research on the impact of these technologies on information warfare and public perception. Additionally, the insights gained can aid security practitioners in developing detection frameworks that are calibrated to the evolving threat landscape posed by generative AI, enhancing preparedness against potential manipulation of public opinion.

Authors: Liz Cho, Dongwook Yoon
Source: arXiv:2605.13785
URL: https://arxiv.org/abs/2605.13785v1