Google announced on Thursday that it will suspend generating photos of people on its generative artificial intelligence (AI) tool Gemini, following “recent issues” with the functionality.
The IT giant launched this new image generation function in its Gemini AI software at the beginning of February, which was simultaneously launched in Canada.
Since then, critics have appeared on social media to point out inaccuracies regarding some of the images created, particularly regarding gender and diversity in historical subjects.
Google did not refer to specific images in its statement, but Internet users noticed this on X (formerly Twitter). twin It appears that white people are not represented in the images created. Among these posts, several netizens posted a request about a German soldier from 1943, under the Nazi regime, which led to images of Asian or black soldiers.
As a former Google X employee told me it was It is difficult to convince Google Gemini to acknowledge the existence of white people
.
We are working to resolve recent issues with Gemini's image creation functionality. In the meantime, we'll be pausing the creation of people photos and will make an improved version of this feature available soon
Jack Krawczyk, Google's product manager responsible for Gemini, explained in a statement sent toFrance Press agency.
Generate images usingAmnesty International Gemini leads to a wide range of people. This is generally a good thing, because people use it around the world. But we failed in this case.
In a reaction published on the X website (formerly Twitter), Elon Musk posted a picture in which he distinguishes between his own tool and Amnesty Internationalabsorbed Maximum search for truth
with Google (twin) And the'OpenAI (ChatGPT), which he describes as racist
And based on I wake up
.
While presenting Gemini to journalists in Canada in early February, Google explained that A big party
The work in basic research and responsible use of artificial intelligence takes place in Canada.
Not the first
This is not the first time Google's AI has been targeted by diversity issues.
About ten years ago, Google Photos identified two black people in a photo as gorillas. The California giant had to apologise.
Google isn't the only company that has been singled out for producing images that include stereotypes. investigating in Washington Post (A new window) In particular, he noted the image generator Stable spread The study showed that food aid recipients are mostly dark-skinned people, while most are white, in the United States.
Since the end of 2022 and the success of ChatGPT, it has becomeAmnesty International Producer and able to produce all types of content (text, sounds, images or videos) upon simple request in everyday language, and arouses great enthusiasm. Tech giants are competing to deploy security tools.Amnesty International For organizations and individuals.
More and more amazing and now able to create 1-minute videos thanks to Sora software fromOpenAI The technology, demonstrated in mid-February, triggered safeguards.
Last week, twenty of the most advanced companies in the field participated, including Meta, Microsoft, Google, OpenAIThus, they announced their commitment to developing new technologies to identify misinformation content using…Amnesty International.
With information from Agence France-Presse
“Hardcore beer fanatic. Falls down a lot. Professional coffee fan. Music ninja.”
More Stories
Word, Excel, and PowerPoint are free
This magnificent crater bears witness to the existence of water ice on Mars
Discovery of the first rocky atmosphere of an exoplanet on Cancri 55 e