← Volver al blog
Investigación IAMedia

Web poisoning: Why training-data provenance matters

Mutable web sources can threaten dataset integrity. Manipulating a dataset and demonstrating a particular model failure are separate steps.

Revisión editorial:

El texto de este artículo está disponible en inglés en esta versión del sitio.

Contexto y fuentes

Poisoning Web-Scale Training Datasets is Practical investigates attacks on the acquisition of web training material. Discussed mechanisms include controlling content retrieved later and modifying publicly editable sources. A list of addresses does not freeze the content returned by those addresses at a future download. [1]

Alcance y límites

A small amount of controllable data does not automatically establish a specific backdoor in every resulting model. Percentages need their experimental context. [1]

Explicación sencilla

A purchasing list records where goods should come from. Yesterday’s supplier address does not by itself guarantee what will be delivered today.

Recomendaciones prácticas

Our recommendation: record retrieval dates, provenance and versions of imported content. Identify sources that can change after dataset creation, including domains that disappear or change ownership. This documentation helps narrow an investigation when outputs become suspicious. Treat integrity as a separate assessment rather than assuming a large or well-known dataset is trustworthy.

Fuente primaria / referencia

  1. [1] Carlini et al. (2023, revised 2024): Poisoning Web-Scale Training Datasets