Abstract
The emergence of coordinated autonomous AI agents is fundamentally altering human belief systems and public discourse at scales and speeds that outpace existing characterization methods. This paper argues that AI swarm intelligence, the coordinated influence of autonomous agents on human populations, represents a phase shift that cannot be understood by simply scaling single-agent analysis. Instead, it must be analyzed as a sociotechnical system where societal impact emerges from the tight coupling of individual agent capabilities, multi-agent coordination, and population dynamics. Because the shift from single- to multi-agent systems corresponds to a psychological transition between single- and multiple-source influence, we propose a unified framework and evaluate these systems across operational, individual, and population-level dynamics. We identify concrete research agendas and call on the community to build the evaluation methods, countermeasures, and governance approaches needed to address the long-term societal impact of these coordinated autonomous AI populations.AI assistance disclosure: Claude (Anthropic) was used solely to polish the language of the manuscript. No other AI assistance was used.