Knowledge Base Regulation Tracker Standards & Frameworks OSS Directory AI Incident Database Research Paper Library Case Studies Society Articles News Jobs About Charter Sign inBecome a Member
Home/Research Paper Library/Poisoning Web-Scale Training Datasets is Practical

Poisoning Web-Scale Training Datasets is Practical

Demonstrated two practical ways to poison web-scale datasets, including buying expired domains referenced by LAION, prompting hash-verification and snapshot fixes.

web-scale poisoninglaionexpired domains

Research question

Can an attacker poison the public datasets used to train foundation models?

Key findings

  • Split-view poisoning: buying expired domains in LAION-400M and other datasets for about $60 lets an attacker control 0.01% of the data - Frontrunning poisoning of Wikipedia snapshots is feasible with predictable timing

Sources

PreviousNot What You've Signed Up For: Compromising Real-World LLM-Integrated Applications with Indirect Prompt InjectionNextHarms from Increasingly Agentic Algorithmic Systems

Contribute

Spot an error, or know a record we should add? Members can propose additions and submit corrections.

Submit researchReport a correction or request removal

Related in Research Paper Library

The Price of Safety: Benign-Case Utility and Token Overhead of Memory-Poisoning Defenses in LLM Agents

Memory-poisoning defenses for LLM agents are typically evaluated by their ability to prevent attacks. However, the traffic they process is rarely adversarial. The cost of implementing a defense is paid with each…

Data poisoning2026

Micro-Collaborative Poisoning: A Distributed Attack on RAG Systems

Retrieval-Augmented Generation (RAG) improves large language models by grounding outputs in external knowledge sources, but this dependency also creates a surface for poisoning attacks. This paper introduces…

Data poisoning2026

PentestChain: A Cost-Aware, MCP-Orchestrated Framework for Automated Penetration Testing with Free-Tier LLMs

AI-driven penetration testing has been demonstrated with premium frontier models such as GPT-4, but the per-engagement token cost makes continuous, automated testing unaffordable for the smaller organisations that…

Data poisoning2026