Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Toxic Ink on Immutable Paper: Content Moderation for Ethereum Input Data Messages (IDMs)

Xihan Xiong, Zhipeng Wang, Qin Wang, William Knottenbelt

TL;DR

This paper tackles the problem of toxic content in Ethereum Input Data Messages (IDMs) by proposing two protocol-level moderation frameworks, BUILDERMOD and USERMOD. It shows that BUILDERMOD incurs high in-block latency, while USERMOD provides lower-latency validation and better scalability, making it more practical for moderation-aware Ethereum environments. The work formalizes system and threat models, details the design frameworks, and evaluates latency and cost implications, including a moderation-fee mechanism to incentivize builders. The findings establish a foundation for semantic governance in decentralized systems and highlight the practical tradeoffs between in-block versus pre-transaction moderation, with implications for policy, security, and economics in Web3 communications.

Abstract

Decentralized communication is becoming an important use case within Web3. On Ethereum, users can repurpose the transaction input data field to embed natural-language messages, commonly known as Input Data Messages (IDMs). However, as IDMs gain wider adoption, there has been a growing volume of toxic content on-chain. This trend is concerning, as Ethereum provides no protocol-level support for content moderation. We propose two moderation frameworks for Ethereum IDMs: (i) BUILDERMOD, where builders perform semantic checks during block construction; and (ii) USERMOD, where users proactively obtain moderation proofs from external classifiers and embed them in transactions. Our evaluation reveals that BUILDERMOD incurs high block-time overhead, which limits its practicality. In contrast, USERMOD enables lower-latency validation and scales more effectively, making it a more practical approach in moderation-aware Ethereum environments. Our study lays the groundwork for protocol-level content governance in decentralized systems, and we hope it contributes to the development of a decentralized communication environment that is safe, trustworthy, and socially responsible.

Toxic Ink on Immutable Paper: Content Moderation for Ethereum Input Data Messages (IDMs)

TL;DR

This paper tackles the problem of toxic content in Ethereum Input Data Messages (IDMs) by proposing two protocol-level moderation frameworks, BUILDERMOD and USERMOD. It shows that BUILDERMOD incurs high in-block latency, while USERMOD provides lower-latency validation and better scalability, making it more practical for moderation-aware Ethereum environments. The work formalizes system and threat models, details the design frameworks, and evaluates latency and cost implications, including a moderation-fee mechanism to incentivize builders. The findings establish a foundation for semantic governance in decentralized systems and highlight the practical tradeoffs between in-block versus pre-transaction moderation, with implications for policy, security, and economics in Web3 communications.

Abstract

Decentralized communication is becoming an important use case within Web3. On Ethereum, users can repurpose the transaction input data field to embed natural-language messages, commonly known as Input Data Messages (IDMs). However, as IDMs gain wider adoption, there has been a growing volume of toxic content on-chain. This trend is concerning, as Ethereum provides no protocol-level support for content moderation. We propose two moderation frameworks for Ethereum IDMs: (i) BUILDERMOD, where builders perform semantic checks during block construction; and (ii) USERMOD, where users proactively obtain moderation proofs from external classifiers and embed them in transactions. Our evaluation reveals that BUILDERMOD incurs high block-time overhead, which limits its practicality. In contrast, USERMOD enables lower-latency validation and scales more effectively, making it a more practical approach in moderation-aware Ethereum environments. Our study lays the groundwork for protocol-level content governance in decentralized systems, and we hope it contributes to the development of a decentralized communication environment that is safe, trustworthy, and socially responsible.

Paper Structure

This paper contains 50 sections, 19 equations, 8 figures, 2 tables.

Table of Contents

  1. Introduction
  2. Ethereum Fundamentals
  3. Input Data Messages (IDMs)
  4. Use of IDMs for Toxic Messaging
  5. Impact of On-chain Toxic IDMs
  6. Challenges for On-chain Moderation of Toxic IDMs
  7. Content Moderation in Web2
  8. Rules
  9. Detection
  10. Enforcement
  11. Content Moderation in Web3
  12. Challenges for On-chain Moderation
  13. Legal Basis
  14. System and Threat Model
  15. System Model
  16. ...and 35 more sections

Figures (8)

  • Figure 1: Overview of Alice's Ethereum transaction lifecycle under the MEV-Boost architecture.
  • Figure 2: Overview of the BuilderMod framework.
  • Figure 3: Overview of the UserMod framework.
  • Figure 4: Comparison of per-transaction validation time ($\mu$s) without and with moderation (BuilderMod). Results are based on outputs from the openai/gpt-4.1 model.
  • Figure 5: Comparison of per-transaction validation time ($\mu$s) without and with moderation (UserMod). Results are based on outputs from the openai/gpt-4.1 model.
  • ...and 3 more figures