https://purl.stanford.edu/kh752sm9123

Stanford University Libraries [sul-logo-924f9a4b30a2] [sul-logo-st]
Feedback
Stanford Digital Repository

Identifying and Eliminating CSAM in Generative ML Training Data and
Models

Show Content
Show Content

Abstract/Contents

Abstract

    Generative Machine Learning models have been well documented as
    being able to produce explicit adult content, including child
    sexual abuse material (CSAM) as well as to alter benign imagery
    of a clothed victim to produce nude or explicit content. In this
    study, we examine the LAION-5B dataset--parts of which were used
    to train the popular Stable Diffusion series of models--to attempt
    to measure to what degree CSAM itself may have played a role in
    the training process of models trained on this dataset. We use a
    combination of PhotoDNA perceptual hash matching, cryptographic
    hash matching, k-nearest neighbors queries and ML classifiers.

    This methodology detected many hundreds of instances of known
    CSAM in the training set, as well as many new candidates that
    were subsequently verified by outside parties. We also provide
    recommendations for mitigating this issue for those that need to
    maintain copies of this training set, building future training
    sets, altering existing models and the hosting of models trained
    on LAION-5B.

Description

Type of resource text
Publication date December 20, 2023

Creators/Contributors

   Author     Thiel,      ORCiD icon https://orcid.org/
              David       0000-0002-0947-5921 (unverified)
Research team Hancock,    ORCiD icon https://orcid.org/
    head      Jeffrey     0000-0001-5367-2677 (unverified)

Subjects

Subject Machine learning
Subject Child abuse
 Genre  Text
 Genre  Report
 Genre  Technical report

Bibliographic information

  DOI    https://doi.org/10.25740/kh752sm9123
Location https://purl.stanford.edu/kh752sm9123

Access conditions

Use and reproduction
    User agrees that, where applicable, content will not be used to
    identify or to otherwise infringe the privacy or confidentiality
    rights of individuals. Content distributed via the Stanford
    Digital Repository may be subject to additional license and use
    restrictions applied by the depositor.
License
    This work is licensed under a Creative Commons Attribution Non
    Commercial No Derivatives 4.0 International license (CC BY-NC-ND)
    .

Preferred citation

Preferred citation
    Thiel, D. (2023). Identifying and Eliminating CSAM in Generative
    ML Training Data and Models. Stanford Digital Repository.
    Available at https://purl.stanford.edu/kh752sm9123. https://
    doi.org/10.25740/kh752sm9123.

Collection

Stanford Internet Observatory, Freeman Spogli Institute for
International Studies

View other items in this collection in SearchWorks

Contact information

Contact
    dthiel@stanford.edu

Also listed in

View in SearchWorks

Loading usage metrics...

Stanford Libraries

  * Hours & locations
  * My Account
  * Ask us
  * System status

Stanford University

  * Stanford Home
  * Maps & Directions
  * Search Stanford
  * Emergency Info

  * Terms of Use
  * Privacy
  * Copyright
  * Trademarks
  * Non-Discrimination
  * Accessibility

(c) Stanford University.   Stanford, California 94305.