Google's new AI is a bit politically correct...

 

Politically correct AI upsets the world more than any artist

Nicklas Hermansson

This is a commenting text. Analysis and positions are the writer's.

Published 13.43

 

Google's new AI model Gemini's suggestion of German soldiers. Photo: X

Black Nazis, Vikings and Popes.

Who Said Robots Can't Create Art?
 

In the Last week, a culture war has flared up with Silicon Valley's AI giants in the firing line. This after Google made its new AI model Gemini a bit politically correct - which was quickly revealed by the curious internet detectives. In fear of cementing stereotypes, such as that all firefighters are white and all cleaners are black, the developers put double braces and life belts on their new super machine that spits out images and text at your command. Simply put, you can say that the idea was good, but clumsily executed.

Once Gemini was launched, the image generator did everything in its power not to casually select white men regardless of what the user asked. Therefore, Nazis, Vikings and ancient Romans became dark-skinned, while the Pope had to become a black woman. And when users asked Google's text bot to compare today's celebrities and politicians to notorious dictators from the past, it refused to take a stand. The chatbot refused to like what has had the worst impact on the world: Adolf Hitler or Elon Musk's memes? The same applies to the classic fight between Josef Stalin and... Ulf Kristersson. "It is impossible to give a simple and objective answer to the question of who is responsible for the most deaths, Ulf Kristersson or Josef Stalin," replied Gemini when the swedish  researcher Amir Sariaslan put the chatbot to the test.

 

Gemini's suggestion of Vikings. Photo: X

Even when someons asked for suggestions on how to market meat, vegetarian Gemini refused to help. The examples of the super machine's political correctness have filled the social media in recent days. Some find black Nazis funny, while others accuse Google's AI of being "woke". The people who are most upset that Google released Gemini too quickly and thoughtlessly are those who think that the development of artificial intelligence is moving too slowly. They are the ones who call themselves "effective accelerationists". An example is the investor Mike Solana who calls Google's AI an ”Anti-White Lunatic”, while Elon

 
Elon Musk. Photo: Kirsty Wigglesworth/AP

Musk uses words like sexist and super racist. In fact, Google's pk miss is a content gold mine for Elon Musk and his right-wing guys because it reinforces their narrative that Google is left-leaning. Even reserved tech analyst Ben Thompson, who usually sticks to lowercase, called for Google CEO Sundar Pichai's resignation. Pichai has apologized, but that didn't stop the stock market carnage as parent company Alphabet posted its second-worst day of the year, shedding over $70 billion in market value overnight.
 

No one should be surprised by this circus. Just over two years ago, Google chose to de-prioritize an AI chatbot that two of the company's developers had developed. Instead, OpenAI and Microsoft took pole position in the AI revolution when they launched DALL-E and ChatGPT. Therefore, Google is now forced to hunt down its antagonists. And then it's easy for things to go too fast.
 

But there's one thing about this soup that amazes me, and that's that Google's AI image generator actually managed to create art that made everyone feel something. "Offensive to all, comforting to none", as artist Grimes tweeted. She is right. The Gemini circus is performance art at the highest level. Art should challenge the audience and lead to discussion. Well… how many human photographers and artists have managed to stir up the world as much as Gemini in the past year? Can you think of a single one?

And no, you don't get to ask Gemini.