Google brings back controversial AI image generation tool after backlash

Google AI has reintroduced the human image generation tool, which was previously disabled for producing historically inaccurate images.

In a blog post on Wednesday, Google announced the upgrade of its Imagen 3 model, which was initially introduced in May. The updated model will soon begin generating images of people for Gemini Advanced, Business, and Enterprise users.

“Over the coming days, we’ll also start to roll out the generation of images of people, with an early access version for our Gemini Advanced, Business, and Enterprise users, starting in English,” the company said.

“With Imagen 3, we’ve made significant progress in providing a better user experience when generating images of people.”

The AI image generator will not support the generation of photorealistic, identifiable individuals, depictions of minors, or excessively gory, violent, or sexual scenes, Google reminded users.

Pexels, Google

The description makes it clear that Google wants to avoid generating controversial images. The search giant clarified that the prohibition on “identifiable individuals” includes “certain queries that could lead to outputs of prominent people.”

Google disabled human image generation in Gemini back in February after people complained it was generating factually and historically inaccurate images. People criticized Gemini’s portrayal of specific figures, such as the US Founding Fathers or groups like Nazi-era German soldiers, as people of color.

People also complained that Gemini won’t generate images of people of their ethnicity. One user tried to generate an image of a “blonde-haired, blue-eyed, Scandinavian woman,” but Gemini produced an image of an “Indian woman with dyed hair.”

At that time, Google apologized for what it described as “inaccuracies in some historical image generation depictions” with Gemini, saying the AI tool’s attempts at generating a “wide range” of images “missed the mark.”

The image generation upgrade comes shortly after Elon Musk acquired X, which also included a similar feature to Grok, and Midjourney opened its image generator to people without Discord accounts. For local image generation, however, Stable Diffusion remains an easy way to locally generate images, if you’re willing to put the effort into it.