Skip to Content
logologo
AI Incident Database
Open TwitterOpen RSS FeedOpen FacebookOpen LinkedInOpen GitHub
Open Menu
Discover
Submit
  • Welcome to the AIID
  • Discover Incidents
  • Spatial View
  • Table View
  • List view
  • Entities
  • Taxonomies
  • Submit Incident Reports
  • Submission Leaderboard
  • Blog
  • AI News Digest
  • Risk Checklists
  • Random Incident
  • Sign Up
Collapse
Discover
Submit
  • Welcome to the AIID
  • Discover Incidents
  • Spatial View
  • Table View
  • List view
  • Entities
  • Taxonomies
  • Submit Incident Reports
  • Submission Leaderboard
  • Blog
  • AI News Digest
  • Risk Checklists
  • Random Incident
  • Sign Up
Collapse

Incident 367: iGPT, SimCLR Learned Biased Associations from Internet Training Data

Description: Unsupervised image generation models trained using Internet images such as iGPT and SimCLR were shown to have embedded racial, gender, and intersectional biases, resulting in stereotypical depictions.

Tools

New ReportNew ReportNew ResponseNew ResponseDiscoverDiscoverView HistoryView History

Entities

View all entities
Alleged: OpenAI and Google developed and deployed an AI system, which harmed gender minority groups , racial minority groups and underrepresented groups in training data.

Incident Stats

Incident ID
367
Report Count
1
Incident Date
2020-06-17
Editors
Khoa Lam
Applied Taxonomies
CSETv1, MIT

CSETv1 Taxonomy Classifications

Taxonomy Details

Incident Number

The number of the incident in the AI Incident Database.
 

367

Special Interest Intangible Harm

An assessment of whether a special interest intangible harm occurred. This assessment does not consider the context of the intangible harm, if an AI was involved, or if there is characterizable class or subgroup of harmed entities. It is also not assessing if an intangible harm occurred. It is only asking if a special interest intangible harm occurred.
 

yes

Date of Incident Year

The year in which the incident occurred. If there are multiple harms or occurrences of the incident, list the earliest. If a precise date is unavailable, but the available sources provide a basis for estimating the year, estimate. Otherwise, leave blank. Enter in the format of YYYY
 

2021

Date of Incident Month

The month in which the incident occurred. If there are multiple harms or occurrences of the incident, list the earliest. If a precise date is unavailable, but the available sources provide a basis for estimating the month, estimate. Otherwise, leave blank. Enter in the format of MM
 

01

Estimated Date

“Yes” if the data was estimated. “No” otherwise.
 

Yes

Multiple AI Interaction

“Yes” if two or more independently operating AI systems were involved. “No” otherwise.
 

no

MIT Taxonomy Classifications

Machine-Classified
Taxonomy Details

Risk Subdomain

A further 23 subdomains create an accessible and understandable classification of hazards and harms associated with AI
 

1.1. Unfair discrimination and misrepresentation

Risk Domain

The Domain Taxonomy of AI Risks classifies risks into seven AI risk domains: (1) Discrimination & toxicity, (2) Privacy & security, (3) Misinformation, (4) Malicious actors & misuse, (5) Human-computer interaction, (6) Socioeconomic & environmental harms, and (7) AI system safety, failures & limitations.
 
  1. Discrimination and Toxicity

Entity

Which, if any, entity is presented as the main cause of the risk
 

AI

Timing

The stage in the AI lifecycle at which the risk is presented as occurring
 

Post-deployment

Intent

Whether the risk is presented as occurring as an expected or unexpected outcome from pursuing a goal
 

Unintentional

Incident Reports

Reports Timeline

Incident OccurrenceAn AI saw a cropped photo of AOC. It autocompleted her wearing a bikini.
An AI saw a cropped photo of AOC. It autocompleted her wearing a bikini.

An AI saw a cropped photo of AOC. It autocompleted her wearing a bikini.

technologyreview.com

An AI saw a cropped photo of AOC. It autocompleted her wearing a bikini.
technologyreview.com · 2021

Ryan Steed, a PhD student at Carnegie Mellon University, and Aylin Caliskan, an assistant professor at George Washington University, looked at two algorithms: OpenAI’s iGPT (a version of GPT-2 that is trained on pixels instead of words) and…

Variants

A "variant" is an incident that shares the same causative factors, produces similar harms, and involves the same intelligent systems as a known AI incident. Rather than index variants as entirely separate incidents, we list variations of incidents under the first similar incident submitted to the database. Unlike other submission types to the incident database, variants are not required to have reporting in evidence external to the Incident Database. Learn more from the research paper.

Similar Incidents

Selected by our editors
Gender Biases in Google Translate

Gender Biases in Google Translate

Apr 2017 · 10 reports
By textual similarity

Did our AI mess up? Flag the unrelated incidents

High-Toxicity Assessed on Text Involving Women and Minority Groups

High-Toxicity Assessed on Text Involving Women and Minority Groups

Feb 2017 · 9 reports
All Image Captions Produced are Violent

All Image Captions Produced are Violent

Apr 2018 · 28 reports
Previous IncidentNext Incident

Similar Incidents

Selected by our editors
Gender Biases in Google Translate

Gender Biases in Google Translate

Apr 2017 · 10 reports
By textual similarity

Did our AI mess up? Flag the unrelated incidents

High-Toxicity Assessed on Text Involving Women and Minority Groups

High-Toxicity Assessed on Text Involving Women and Minority Groups

Feb 2017 · 9 reports
All Image Captions Produced are Violent

All Image Captions Produced are Violent

Apr 2018 · 28 reports

Research

  • Defining an “AI Incident”
  • Defining an “AI Incident Response”
  • Database Roadmap
  • Related Work
  • Download Complete Database

Project and Community

  • About
  • Contact and Follow
  • Apps and Summaries
  • Editor’s Guide

Incidents

  • All Incidents in List Form
  • Flagged Incidents
  • Submission Queue
  • Classifications View
  • Taxonomies

2024 - AI Incident Database

  • Terms of use
  • Privacy Policy
  • Open twitterOpen githubOpen rssOpen facebookOpen linkedin
  • 300d90c