Coinbase Rolls Out AI-Driven ERC-20 Scam Token Detection System

blockchain.news:

Coinbase's Chief Legal Officer, Paul Grewal, recently shed light on how their Engineering team is harnessing Artificial Intelligence (AI) to root out ERC-20 scam tokens.This innovation was disclosed on October 6, 2023, by Yifan Xu, Indra Rustandi, Yao Ma, and Vijay Dialani from the engineering team at Coinbase. The innovative ERC-20 Scam Token Detection System is a blend of smart contract auditing and machine learning prediction aimed at identifying both known and emergent scam types, marking a significant leap towards ensuring a safer crypto space.

The burgeoning realm of cryptocurrency is not without its share of scams, especially surrounding new and unverified tokens. Fraudsters employ a variety of devious tactics, ranging from Honeypot scams, which are deceptive traps to ensnare investors, to Internal Fees scams involving hidden or unusually high transaction fees. The evolving nature of these scams presents a continuous challenge, with new scam types emerging daily, posing a significant threat to both investors and the broader crypto ecosystem.

To combat these challenges, Coinbase has developed the Scam Token Detection System which employs a two-pronged strategy: 

Smart contract auditing is a proactive measure to identify and filter out known scam types. By meticulously examining the integrity of tokens, this step helps to mitigate the risk of fraud by excluding tokens associated with known malicious activities, capturing and cataloging them for future reference.

On the flip side, the system employs a machine learning framework to detect unknown scam types by identifying abnormal activity patterns. For instance, unusual patterns in time-series transactions among a concentrated group of accounts could be