Incidente 367: iGPT y SimCLR aprendieron asociaciones sesgadas a partir de datos de entrenamiento de Internet
Descripción: Se demostró que los modelos de generación de imágenes no supervisados entrenados utilizando imágenes de Internet, como iGPT y SimCLR, tenían sesgos raciales, de género e interseccionales incorporados, lo que daba como resultado representaciones estereotipadas.
Entidades
Ver todas las entidadesPresunto: un sistema de IA desarrollado e implementado por OpenAI y Google, perjudicó a gender minority groups , racial minority groups y underrepresented groups in training data.
Clasificaciones de la Taxonomía CSETv1
Detalles de la TaxonomíaIncident 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
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.
- 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
Informes del Incidente
Cronología de Informes
Ryan Steed, estudiante de doctorado en la Universidad Carnegie Mellon, y Aylin Caliskan, profesora asistente en la Universidad George Washington, analizaron dos algoritmos: [iGPT de OpenAI](https://www.technologyreview.com/2020/07/16/100528…
Variantes
Una "Variante" es un incidente de IA similar a un caso conocido—tiene los mismos causantes, daños y sistema de IA. En lugar de enumerarlo por separado, lo agrupamos bajo el primer incidente informado. A diferencia de otros incidentes, las variantes no necesitan haber sido informadas fuera de la AIID. Obtenga más información del trabajo de investigación.
¿Has visto algo similar?
Incidentes Similares
Selected by our editors
Gender Biases in Google Translate
· 10 informes
Did our AI mess up? Flag the unrelated incidents
All Image Captions Produced are Violent
· 28 informes
Incidentes Similares
Selected by our editors
Gender Biases in Google Translate
· 10 informes
Did our AI mess up? Flag the unrelated incidents
All Image Captions Produced are Violent
· 28 informes