Following a temporary suspension earlier this year, Google is reintroducing the ability to generate images of people through its Gemini AI chatbot. An early access version of this capability will be rolled out to Gemini Advanced, Business, and Enterprise users in English over the coming days.
This upgrade is powered by Imagen 3, the latest iteration of Google’s AI text-to-image generator, which was launched through the AI Test Kitchen earlier this month. Imagen 3 is capable of generating a wide range of images, from photorealistic landscapes to textured oil paintings, using brief descriptions.
The feature was initially paused in February due to concerns over historically inaccurate images, including racially diverse Nazis. However, the upgraded Imagen 3 model features built-in safeguards and has been tested to perform favorably compared to other image generation models.
It is important to note that this feature is distinct from Google’s new Remagine feature, which allows users to incorporate AI elements into photos taken on Pixel 9 phones. Additionally, Gemini will not permit users to create photorealistic images of public figures, content involving minors, or explicit scenes.
While not all images generated by Gemini will be perfect, Google will continue to gather feedback from early access users to improve the feature. The company plans to expand this capability to more users and languages in the near future.”