Google
was forced to recalculate its strategy in the race for dominance of generative artificial intelligence.
The internet giant announced the
interruption of its function to create images
after the Gemini (formerly Bard) platform
represented German troops from the Nazi era as people of color
, among other criticisms received in recent days.
The American technology giant had launched a new improved version of its Gemini program
, known as Ultra 1.0 , on February 8 in some countries
, but now acknowledged that it will have to "fix recent problems" with the
imaging function
.
"We are suspending the generation of images of people and will soon re-release an improved version," the company said in a statement, the
AFP
news agency reported .
Two days ago, a user of
The cover of the New York Post newspaper shows a black George Washington.
The AI had generated four images of soldiers: one was white, one was black, and
two were women of color
, according to X user named John L.
"@GoogleAI has an added
diversity mechanism
that someone didn't design very well or test," he wrote.
Another who joined the criticism about Gemini's performance was
Elon Musk
himself , owner of X, who attached in his defense an alternative version of a
black George Washington
.
"
The
woke
virus is killing Western civilization
. Google does the same with its search results. Facebook and Instagram too. And Wikipedia," the tycoon shot.
Complaints from Gemini users on the social network X.
There were even users who directly called
Google's chatbot
racist by asking it to generate images of "beautiful trans women."
Along the same lines, Gemini delivered
only photographs with characters of color
.
Tech companies see AI as the future for everything from search engines to smartphone cameras.
But AI programs, not just those produced by Google, were widely criticized for
perpetuating racial bias
in their results.
This is what the soldiers of Nazi Germany looked like according to Gemini.
Big tech companies were often accused of releasing AI products before they were properly tested.
Google has a
spotty track record
in launching AI products, at a time when advances by rivals Microsoft Corp. and OpenAI pose a potential threat to its core Internet search business.
The company apologized a year ago after an ad for its newly launched Bard chatbot showed the program
incorrectly answering
a basic question about astronomy.
Steady advances in recent times in AI have raised concerns about the possibility that the technology allows for
falsification, misinformation and bias in its results.
The updated AI model, called Gemini 1.5 Pro, was made available last week to customers and cloud developers so they can test its new features and eventually create new business applications.
Both Google and its rivals are estimated
to have invested billions of dollars
in improving their generative AI capabilities and seeking to attract corporate customers to show that their investments are paying off.
What is Gemini and how does it work?
Gemini, Google's new AI.
Google Photo
Google surprised at the beginning of 2024 with the announcement of the update of its artificial intelligence Bard, which it renamed Gemini, a year after its official launch.
ChatGPT's rival chatbot uses machine learning techniques to perform various tasks, such as speech recognition, natural language processing, and decision making.
Gemini is based on a deep neural network and uses large amounts of data to improve its accuracy and efficiency in its tasks.
Its goal is to improve the interaction between humans and computers, allowing for more natural and efficient communication.
Currently, the technology giant uses it in several Google products, such as the virtual assistant Google Assistant and the translation platform Google Translate.
It also powers the
new artificial intelligence functions
of the brand new Samsung Galaxy S24 cell phone series.
SL