Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Stop the Nonconsensual Use of Nude Images in Research

Princessa Cintaqia, Arshia Arya, Elissa M Redmiles, Deepak Kumar, Allison McDonald, Lucy Qin

TL;DR

This position paper argues that a substantial portion of nudity-detection research relies on nonconsensually collected nude images, contributing to image-based sexual abuse (IBSA). It conducts a systematic review of 150 papers across venues, revealing pervasive nonconsensual data collection, inadequate consent consideration, and distribution of nude content via publications, annotations, and datasets. The authors advocate ethical realignment toward consent, data governance, and stakeholder-aligned data sources, including a participatory data-trust model and cautious use of AI-generated data as a harm-mitigation option. They call on publishing venues to strengthen norms and reviewers to enforce ethics guidelines to balance research objectives with dignity and safety of individuals depicted.

Abstract

In order to train, test, and evaluate nudity detection models, machine learning researchers typically rely on nude images scraped from the Internet. Our research finds that this content is collected and, in some cases, subsequently distributed by researchers without consent, leading to potential misuse and exacerbating harm against the subjects depicted. This position paper argues that the distribution of nonconsensually collected nude images by researchers perpetuates image-based sexual abuse and that the machine learning community should stop the nonconsensual use of nude images in research. To characterize the scope and nature of this problem, we conducted a systematic review of papers published in computing venues that collect and use nude images. Our results paint a grim reality: norms around the usage of nude images are sparse, leading to a litany of problematic practices like distributing and publishing nude images with uncensored faces, and intentionally collecting and sharing abusive content. We conclude with a call-to-action for publishing venues and a vision for research in nudity detection that balances user agency with concrete research objectives.

Stop the Nonconsensual Use of Nude Images in Research

TL;DR

This position paper argues that a substantial portion of nudity-detection research relies on nonconsensually collected nude images, contributing to image-based sexual abuse (IBSA). It conducts a systematic review of 150 papers across venues, revealing pervasive nonconsensual data collection, inadequate consent consideration, and distribution of nude content via publications, annotations, and datasets. The authors advocate ethical realignment toward consent, data governance, and stakeholder-aligned data sources, including a participatory data-trust model and cautious use of AI-generated data as a harm-mitigation option. They call on publishing venues to strengthen norms and reviewers to enforce ethics guidelines to balance research objectives with dignity and safety of individuals depicted.

Abstract

In order to train, test, and evaluate nudity detection models, machine learning researchers typically rely on nude images scraped from the Internet. Our research finds that this content is collected and, in some cases, subsequently distributed by researchers without consent, leading to potential misuse and exacerbating harm against the subjects depicted. This position paper argues that the distribution of nonconsensually collected nude images by researchers perpetuates image-based sexual abuse and that the machine learning community should stop the nonconsensual use of nude images in research. To characterize the scope and nature of this problem, we conducted a systematic review of papers published in computing venues that collect and use nude images. Our results paint a grim reality: norms around the usage of nude images are sparse, leading to a litany of problematic practices like distributing and publishing nude images with uncensored faces, and intentionally collecting and sharing abusive content. We conclude with a call-to-action for publishing venues and a vision for research in nudity detection that balances user agency with concrete research objectives.
Paper Structure (30 sections, 1 figure)

This paper contains 30 sections, 1 figure.

Table of Contents

  1. Introduction
  2. Related Work
  3. Methods
  4. Data Collection and Distribution Practices
  5. Using nude images nonconsensually in research is harmful
  6. Subjects' lack of consent is unacknowledged.
  7. Using commercial content is still harmful.
  8. Nude images are collected in excess.
  9. Researchers are amassing datasets of IBSA
  10. Nonconsensual intimate content is difficult to remove from the Internet.
  11. Abuse material is incidentally collected.
  12. Abuse material is intentionally collected.
  13. Data sources are largely undisclosed.
  14. Images are distributed without consent via publication, annotation, and open science
  15. Research Practices
  16. ...and 15 more sections

Figures (1)

  • Figure 1: Number of papers included by 5-year range in the set of 150 included papers. The stacked bars differentiate between A* papers (pink) and other venues (blue). Note that there were no papers from the years 2000 or 2001 in our dataset.