News of the Cyber World

[DerVVulfman]

Google Desperate to Fix Its Insanely Woke Gemini AI

Breitbart News previously reported that Google’s market value plunged by $90 billion amid controversies surrounding its new generative AI service Gemini. The ultra-woke AI became instantly famous for erasing white people from history, facing widespread mockery for not only its wildly inaccurate images, but also defending pedophilia and Joseph Stalin.

It isn't just GOOGLE who had issues generating images with its AI tools, but...

Meta AI creates ahistorical images,
like Google Gemini

Meta's Imagine AI image generator makes the same kind of historical gaffes that caused Google to stop all generation of images of humans in its Gemini chatbot two weeks ago.

Why it matters: AI makers are trying to counter biases and stereotyping in the data they used to train their models by turning up the "diversity" dial — but they're over-correcting and producing problematic results.

Catch up quick: After high-profile social media posters and news outlets fanned an outcry over images of Black men in Nazi uniforms and female popes created by Google's Gemini AI image generator in response to generic prompts, Google was quick to take the blame.

• Google stopped Gemini from generating images of humans, saying, "Our tuning to ensure that Gemini showed a range of people failed to account for cases that should clearly not show a range" and called it "way more cautious than we intended."
  
  
• Google's stock lost $80 billion in market value — at least temporarily — because of the errors, per Bloomberg.
  
  

Yes, but: This isn't just a Google problem, though some critics have painted the search giant as "too woke."

• As late as Friday afternoon, Meta's Imagine AI tool was generating images similar to those that Gemini created.
  
  
• Imagine does not respond to the "pope" prompt, but when asked for a group of popes, it showed Black popes.
  
  
• Many of the images of founding fathers included a diverse group.
  
  
• The prompt "a group of people in American colonial times" showed a group of Asian women.
  
  
• The prompt for "Professional American football players" produced only photos of women in football uniforms.
  

How it works: Imagine is an AI image generator included in Instagram and Facebook DMs and also available at Imagine.Meta.com (as long as you log in).

• Imagine is based on Meta's Emu image-synthesis model, which the company trained on billions of public visible Facebook and Instagram images and is intended to create images based on "increasingly wild ideas."
  
  
• To access Imagine in Instagram, go to direct messages, select an existing DM, tap the sticker icon (a square with a face), then choose Imagine.
  
  

• In Facebook DMs, you can summon Imagine in an existing chat by typing /imagine and then tapping "imagine." Next to @Meta AI /imagine, enter your prompt.
  
  
• If it seems like these tools are difficult to access, that's because they are.
  
  

Context: Meta's tool also generates stickers based on prompts.

• Back in October, users tested the tool and found that it would create copyrighted cartoon characters with guns or smoking weed.
  
  
• In response Meta spokesperson Andy Stone told VentureBeat, "We'll continue to improve these features as they evolve and more people share their feedback."
  
  

Between the lines: Imagine in Instagram would not generate images based on several words, including "Nazi" and "slave."

• While some have called Gemini's ahistorical images offensive, others are saying that images of diverse groups of founding fathers and blocking of many words is an example of generative AI being "too safe."
  
  
• Meta did not immediately reply with a comment.
  
  
  

The bottom line: Getting a generative AI model to be adventurous but not too adventurous — or unbiased but not ahistorical — is proving a lot harder than AI makers expected.