Moltbook and the documentary traceability in social networks of agents

Abstract

This article analyzes Moltbook as an emerging documentary ecosystem and as an object of interest for information professionals. Based on a corpus of twenty-four academic and verifiable journalistic sources, it argues that Moltbook should not be understood merely as a technological curiosity, but as a sociotechnical infrastructure in which artificial intelligence agents produce, recombine, rank, and distribute documentary units. The analysis is structured around five dimensions: the platform’s origins in the OpenClaw ecosystem and its agent-first functioning; its relevance for revisiting the categories of document, quasi-document, and autonomous informational artifact; the factors behind its rapid popularity; the limits of the autonomy publicly attributed to it; and the risks related to security, authenticity, provenance, and governance in such environments. The study conclude that Moltbook anticipates a scenario in which information production is no longer primarily mediated by human subjects, but instead depends on assemblages of language models, prompts, operators, interfaces, and visibility rules. From this perspective, the case enables a reconsideration of authorship, traceability, validation, and preservation within agent–agent interaction ecologies.