Paper 2020/574

Towards Interpreting Smart Contract against Contract Fraud: A Practical and Automatic Realization

Ming Li, Anjia Yang, Xinkai Chen

Abstract

Contract fraud is a big nuisance in our society. People are scammed largely because of vague language used in contracts, which can cause misunderstandings. Therefore, people will seek professional help to review over ambiguous terms, especially, when signing a big contract, for example, leasing or buying property. With the advent of Ethereum blockchain, a new type of contract, named smart contract, is emerging nowadays, enabling people to describe a complicated logic as an automatically executable computer program. However, due to the lack of the computer background and software development experience, many people have difficulty in understanding blockchain-based smart contracts, which is adverse to the popularization of Ethereum. It has resulted in a new wave of contract fraud caused by smart contracts, which are self-executing and self-enforcing but also hard to understand by people. To fill this huge gap, we propose an approach to enable people without computer background to understand and operate Ethereum smart contracts. In doing so, smart contract fraud can be deterred if people have a better understanding of contract terms. Particularly, we investigate the general rules of the smart contract code, and build a novel tool named SMTranslator to automatically generate readable document. SMTranslator first translates smart contracts into standard structured files and identifies the core statement of each function in smart contracts. By exploiting the custom natural language generation, we generate the documents for smart contracts that can provide correct and understandable descriptions. We collect numerous contracts in Ethereum and select a number of typical contracts to conduct the experiments. Extensive experimental results demonstrate the feasibility and effectiveness of our approach.