Google has decided to bring back to life a feature that was highly anticipated by users: the generation of images of people via its AI chatbot, Gemini. After temporarily suspending this feature due to some controversy related to the generation of historically inaccurate images, the Mountain View giant has announced the return of this feature, but with important improvements and safeguards.
With the integration of Imagen 3, Google’s latest AI image generator, Gemini can now create a wide range of images, from photorealistic landscapes to oil paintings, starting with simple text descriptions. This new version of Gemini, which will be available as a preview for Gemini Advanced, Business, and Enterprise users in the coming days, promises to offer an even richer and more personalized creative experience.
Gemini’s ability to generate images of people has previously raised concerns about the creation of false or offensive content. To address these concerns, Google has implemented a number of security measures within Imagen 3. The model has been trained on a larger and more diverse dataset, and new techniques have been introduced to mitigate the risk of generating inappropriate images.
Despite the improvements made, it is important to note that Gemini is not perfect. Google acknowledges that some of the images generated may be imperfect and encourages users to provide feedback to help the model continue to improve. In addition, some restrictions have been introduced to prevent misuse of the technology, such as prohibiting the generation of images of public figures, minors, or violent content.
The update comes just days after the controversy over the changes to images taken with the Reimagine function (not to be confused with Imagen 3), given that the images created by the AI with the Pixel 9 were a bit too extreme.
What do you think? Are you in favor of reintroducing this feature? Let us know in the comments below.
#Creating #images #people #Gemini #Imagen #problems