Skip to Content
logologo
AI Incident Database
Open TwitterOpen RSS FeedOpen FacebookOpen LinkedInOpen GitHub
Open Menu
Descubrir
Enviar
  • Bienvenido a la AIID
  • Descubrir Incidentes
  • Vista espacial
  • Vista Tabular
  • Vista de lista
  • Entidades
  • Taxonomías
  • Enviar Informes de Incidentes
  • Ranking de Reportadores
  • Blog
  • Resumen de noticias de IA
  • Control de Riesgos
  • Incidente aleatorio
  • Registrarse
Colapsar
Descubrir
Enviar
  • Bienvenido a la AIID
  • Descubrir Incidentes
  • Vista espacial
  • Vista Tabular
  • Vista de lista
  • Entidades
  • Taxonomías
  • Enviar Informes de Incidentes
  • Ranking de Reportadores
  • Blog
  • Resumen de noticias de IA
  • Control de Riesgos
  • Incidente aleatorio
  • Registrarse
Colapsar

Problema 3788

Incidentes Asociados

Incidente 64535 Reportes
Seeming Pattern of Gemini Bias and Sociotechnical Training Failures Harm Google's Reputation

Loading...
After Google Gemini, Facebook parent Meta’s AI tool creates ‘historically inaccurate images
timesofindia.indiatimes.com · 2024

Google Gemini and Meta's Imagine AI faced backlash for generating historically inaccurate images. The models showed people of Southeast Asia as American colonials and people of color as Founding Fathers. The issue highlighted bias and over correction in AI models, raising concerns about the accuracy and diversity of AI-generated content.

Google Gemini faced heat when it returned ‘historically inaccurate’ images, prompting the company to temporarily cease the ability of the AI chatbot to generate photos. Company CEO Sundar Pichai called it ‘unacceptable’ with the tech giant saying that it will release a fine-tuned model. Now, a report has claimed that Facebook parent Meta’s AI tool has created ahistorical images.

According to a report by Axios, Meta's Imagine AI image generator has also returned the similar kind of historical gaffes as Gemini.

A screenshot shared in the report shows AI-generated images from Meta AI's Imagine tool inside Instagram direct messages. In one of the images, “a group of people in [an] American colonial” is depicted by people of Southeast Asia. In another image, the Founding Fathers [of America] are people of colour.

The prompt for “Professional American football players” produced only photos of women in football uniforms, the report said.

Additionally, Imagine AI model does not respond to the “pope” prompt, but when asked for a group of popes, it showed Black popes, it said.

What is the problem

People have taken offence by the images generated by AI models as AI makers claim that it was due to the inclusion of “diversity”, there is a bias and stereotyping in the data they used to train their models. They also say that while diversity is a good thing, their AI models are over-correcting and producing problematic results.

Google’s Gemini created Black men in Nazi uniforms and female popes.

“Our tuning to ensure that Gemini showed a range of people failed to account for cases that should clearly not show a range” Google said, adding that the model is “way more cautious than we intended.”

Leer la Fuente

Investigación

  • Definición de un “Incidente de IA”
  • Definición de una “Respuesta a incidentes de IA”
  • Hoja de ruta de la base de datos
  • Trabajo relacionado
  • Descargar Base de Datos Completa

Proyecto y Comunidad

  • Acerca de
  • Contactar y Seguir
  • Aplicaciones y resúmenes
  • Guía del editor

Incidencias

  • Todos los incidentes en forma de lista
  • Incidentes marcados
  • Cola de envío
  • Vista de clasificaciones
  • Taxonomías

2024 - AI Incident Database

  • Condiciones de uso
  • Política de privacidad
  • Open twitterOpen githubOpen rssOpen facebookOpen linkedin
  • e1b50cd