Problem
Research questions and friction points this paper is trying to address.
Detecting wash trading in NFT markets using AI.
Evaluating manipulation in NFT trades across major exchanges.
Improving accuracy of wash trading detection methods.
Can AI Detect Wash Trading? Evidence from NFTs
🤖 AI Summary
This paper addresses the critical limitation in NFT market wash trading detection—its reliance on indirect statistical proxies or private data, lacking direct on-chain identification capability. We propose the first end-to-end AI estimation framework leveraging only publicly available on-chain data. Methodologically, we innovatively embed classical statistical tests—such as trailing-digit integer regression—into a machine learning pipeline, integrating on-chain behavioral modeling, cross-exchange heterogeneous data calibration, and ensemble-based regression feature engineering. Empirical evaluation across major NFT marketplaces reveals that approximately 38% of transaction volume and 60% of trading value exhibit strong evidence of manipulation. Our algorithm substantially reduces both exchange-level and transaction-level false positive rates, demonstrating strong generalizability and robustness not only for NFTs but also across broader cryptocurrency asset classes.
📝 Abstract
Existing studies on crypto wash trading often use indirect statistical methods or leaked private data, both with inherent limitations. This paper leverages public on-chain NFT data for a more direct and granular estimation. Analyzing three major exchanges, we find that ~38% (30-40%) of trades and ~60% (25-95%) of traded value likely involve manipulation, with significant variation across exchanges. This direct evidence enables a critical reassessment of existing indirect methods, identifying roundedness-based regressions `a la Cong et al. (2023) as most promising, though still error-prone in the NFT setting. To address this, we develop an AI-based estimator that integrates these regressions in a machine learning framework, significantly reducing both exchange- and trade-level estimation errors in NFT markets (and beyond).
Innovation
Methods, ideas, or system contributions that make the work stand out.
Uses public on-chain NFT data
Develops AI-based estimator
Integrates regressions in ML framework
🔎 Similar Papers
No similar papers found.
