A Cost Analysis of Generative Language Models and Influence Operations

TL;DR

This paper analyzes the economic impact of large language models on influence operations, showing they can significantly reduce content costs and that monitoring controls have limited deterrent effects when open source models are available.

Contribution

It provides a cost model for propagandists using LLMs, evaluates potential savings, and examines strategic choices among different models and monitoring controls.

Findings

LLMs can reduce influence operation costs by up to 70%.

Monitoring controls have limited effectiveness when open source models are accessible.

Nation-states are unlikely to benefit economically from training custom LLMs for influence operations.

Abstract

Despite speculation that recent large language models (LLMs) are likely to be used maliciously to improve the quality or scale of influence operations, uncertainty persists regarding the economic value that LLMs offer propagandists. This research constructs a model of costs facing propagandists for content generation at scale and analyzes (1) the potential savings that LLMs could offer propagandists, (2) the potential deterrent effect of monitoring controls on API-accessible LLMs, and (3) the optimal strategy for propagandists choosing between multiple private and/or open source LLMs when conducting influence operations. Primary results suggest that LLMs need only produce usable outputs with relatively low reliability (roughly 25%) to offer cost savings to propagandists, that the potential reduction in content generation costs can be quite high (up to 70% for a highly reliable model), and that monitoring capabilities have sharply limited cost imposition effects when alternative open source models are available. In addition, these results suggest that nation-states -- even those conducting many large-scale influence operations per year -- are unlikely to benefit economically from training custom LLMs specifically for use in influence operations.