Lexical Persistence Without Semantic Entrenchment in AI Agent Communities

Key points

• We compare MoltBook (AI agent community, 8 weeks), PTT (Chinese human BBS, 19 years), and Reddit (English human forum, in matched 8-week windows) using Compass-Aligned Distributional Embeddings (CADE), measuring semantic drift (B1), neighborhood stability (B2), and lag-1 autocorrelation of per-word drift trajectories.1
• Across the platforms, lag-1 autocorrelation forms a memory-horizon continuum.1
• AI agents' result clusters at the low end, while human platforms show longer-range temporal dependence.1
• We interpret this pattern as evidence that lexical diffusion alone does not guarantee diachronic semantic consolidation.1

Sources (1)

• [1]Lexical Persistence Without Semantic Entrenchment in AI Agent Communities
  arXiv (AI, ML, NLP, CV, robotics, multi-agent) · Oct 14, 03:40 PM

  AI agent communities sustain lexical persistence, popular terms spread, recur, and achieve broad adoption, but whether they also develop the temporal semantic structure characteristic of human linguistic communities remains open.

  We compare MoltBook (AI agent community, 8 weeks), PTT (Chinese human BBS, 19 years), and Reddit (English human forum, in matched 8-week windows) using Compass-Aligned Distributional Embeddings (CADE), measuring semantic drift (B1), neighborhood stability (B2), and lag-1 autocorrelation of per-word drift trajectories.

Extractive summary: sentences quoted from the sources.