README.md

Code for paper: AI agents reshape consensus formation in human groups

This repository contains the public figure-reproduction code and processed data for the Human-AI consensus manuscript.

Layout

• Top-level *.py: scripts and helper modules for manuscript figures and tables.
• processed_data/: anonymized processed data used by the scripts.
• figures/: created when scripts are run; generated outputs are not committed.

Requirements

The code was tested with the following Python packages:

Package	Version
matplotlib	3.10.5
modelscope	1.26.0
nltk	3.9.2
numpy	2.2.6
openpyxl	3.1.5
pandas	2.2.3
scikit-learn	1.7.1
scipy	1.15.3
seaborn	0.13.2
spacy	3.8.11
statsmodels	0.14.5
torch	2.6.0
transformers	4.51.3
wordcloud	1.9.6

Install the dependencies with:

pip install -r requirements.txt
python -m spacy download en_core_web_sm

The word-cloud script also uses the NLTK WordNet corpus.

Main Outputs

Manuscript item	Script	Output
Fig. 1b	analyze_convergence_new.py	figures/fig1b.pdf
Fig. 1c	analyze_human_agent_separately.py	figures/fig1c.pdf
Fig. 2a-b	run_contribution_analysis_continuous.py	figures/fig2a.pdf, figures/fig2b.pdf
Fig. 2c	plot_fig2_boxplot_v5.py, plot_fig2_wordcloud_panel.py	figures/fig2c_boxplot.pdf, figures/fig2c_wordcloud.pdf
Fig. 3a-c	plot_adoption_persistence_dynamics_by_pair_type.py	figures/fig3a.pdf, figures/fig3bc.pdf
Fig. 3d-g	plot_fig3_combined_continuous.py	figures/fig3de.pdf, figures/fig3fg.pdf
Fig. 4a	plot_subjective_agreement_and_ai_leadership.py	figures/fig4a.pdf
Fig. 4b-d and Supplementary Table 1	plot_ai_penalty_with_controls_with_centrality_distance.py	figures/fig4b.pdf, figures/fig4c.pdf, figures/fig4d.pdf, figures/table_s1.tex