Topics

up-to-the-minute

AI

Amazon

Article image

Image Credits:Kenzo Tribouillard / AFP / Getty Images

Apps

Biotech & Health

mood

Sundar Pichai

Image Credits:Kenzo Tribouillard / AFP / Getty Images

Cloud Computing

Commerce

Crypto

enterprisingness

EVs

Fintech

Fundraising

gizmo

Gaming

Google

Government & Policy

Hardware

Instagram

Layoffs

Media & Entertainment

Meta

Microsoft

privateness

Robotics

Security

Social

Space

Startups

TikTok

Transportation

Venture

More from TechCrunch

event

Startup Battlefield

StrictlyVC

Podcasts

Videos

Partner Content

TechCrunch Brand Studio

Crunchboard

Contact Us

Google suppose it ’s temporarily freeze the ability of Gemini , its flagship productive AI entourage of models , to generate image of people while it work on updating the engineering to improve the historical accuracy of output require delineation of humans .

In aposton the societal media platform X , the caller announced what it cast as a “ break ” on generate image of masses — write that it ’s mould to address “ recent exit ” link up to diachronic inaccuracy .

“ While we do this , we ’re going to hesitate the image genesis of people and will re - release an improved version presently , ” it added .

We ’re already working to address recent issues with Gemini ’s image generation feature . While we do this , we ’re going to hesitate the image generation of people and will re - release an improved reading soon.https://t.co/SLxYPGoqOZ

— News from Google ( @NewsFromGoogle)February 22 , 2024

Google found the Gemini image generation tool before this month . However examples of it generating incongruous images of historical citizenry have been find their wayonto social mediain late Clarence Shepard Day Jr. — such as double of the U.S. Founding Fathers depict as American Indian , dark or Asiatic — run to literary criticism and even guy .

Join us at TechCrunch Sessions: AI

Exhibit at TechCrunch Sessions: AI

Writing in a mail onLinkedIn , Paris - based venture capitalist Michael Jackson joined the cumulation - on today — branding Google ’s AI as “ a ludicrous DEI parody ” . ( DEI stand for ‘ Diversity , Equity and Inclusion . ’ )

In apost on go yesterday , Google confirmed it was “ mindful ” the AI was produce “ inaccuracies in some historic image generation depictions ” , adding in a statement : “ We ’re working to better these kinds of word-painting forthwith . Gemini ’s Al image generation does father a wide range of citizenry . And that ’s broadly speaking a practiced thing because citizenry around the populace habituate it . But it ’s missing the mark here . ”

Generative AI tools produce outputs base on training datum and other parameter , such as model weights .

Such tools have more often present literary criticism for producing outputs that are bias in more stereotypical manner — such asovertly sexualize imagery of womenor by answer to prompt forhigh condition job roles with imagery of snowy human beings .

An early AI range of a function classification tool made by Google caused scandalization , back in 2015 , when it misclassified Black men as gorillas . The company promised to fix the outcome but , asWiredreported a few twelvemonth later , its ‘ fix ’ was saturated workaround : With Google simply block the technical school from recognizing Gorilla gorilla at all .

https://techcrunch.com/2024/02/16/what-is-google-gemini-ai/