arifmohamedkhan/ai-writing-evolutionary-dynamics

Evolutionary Dynamics of AI-Mediated Scientific Writing

Complete experimental logs and reproduction package.

Author: Arif Mohamed Khan Rabi Ahamad
Affiliation: School of Information Studies, Syracuse University
Contact: arabiaha@syr.edu | ORCID: 0009-0001-0986-7570

Contents

Directory	Files	Description
logs/	0	Complete stdout from all experiments
figures/	91	All paper figures (PDF + PNG)
data/	17	Derived datasets (QTE matrices, Price components)
scripts/	0	All analysis code (Google Colab compatible)

Key Claims and Evidence

#	Claim	Value	Log File
1	QTE peaks at τ ≈ 0.35	19/19 fields	logs/01_structural_qte.log
2	Signal rank = 4	Gavish-Donoho	logs/02_math_perspectives.log
3	90% shape change	Wasserstein	logs/02_math_perspectives.log
4	Price → Mode 2: ρ = -0.76	p < 0.001	logs/04_price_equation.log
5	Single-control unstable	ρ = -0.02	logs/06_sensitivity_control.log
6	Shared Mode 2 ↔ Price: ρ = -0.88	p < 0.001	logs/07_multi_control_decomposition.log
7	6-control: ρ = +0.75	p = 0.005	logs/08_robustness.log
8	Classifier-robust	ρ ∈ [0.75, 0.78]	logs/08_robustness.log
9	19/19 fields diversify	0 homogenized	logs/05_price_decomposition.log
10	bioRxiv replication	1.65× ratio	logs/08_robustness.log
11	Adaptive radiation	8/8 decelerate	logs/08_robustness.log
12	Impossibility witness	53% artifact	logs/07_multi_control_decomposition.log

Verification

# Clone this repo
git clone https://huggingface.co/datasets/arabiaha/ai-writing-evolutionary-dynamics
cd ai-writing-evolutionary-dynamics

# Run automated verification (checks all claims against logs)
python scripts/verify_claims.py

Full Reproduction

All scripts run on Google Colab (free tier). Total runtime ~4 hours. Data sources: arXiv (arxiv.org), Semantic Scholar (api.semanticscholar.org), bioRxiv (api.biorxiv.org).

Citation

@article{ahamad2026evolutionary,
  title={Evolutionary Dynamics of AI-Mediated Scientific Writing},
  author={Ahamad, Arif Mohamed Khan Rabi},
  journal={SN Social Sciences},
  year={2026}
}