Prime Highlights:
- Google revamps its AI-generated images with the introduction of Imagen 3 in Gemini, offering enhanced visual realism and more user control.
- Generating images of people has been resumed with solid safety and accuracy standards.
Key Facts:
- Imagen 3 enables users to create a range of image styles, from photorealistic landscapes to art-based depictions.
- The feature is now accessible to Gemini Advanced, Business, and Enterprise users, with English prompt support.
Key Background:
Google’s Gemini AI was making headlines in early 2024 for producing controversial and historically inaccurate portraits of people, leading to public outrage. In response, Google temporarily suspended the feature utilized to produce images of people and promised to review its content policies, system controls, and evaluation processes.
In the following months, Google overhauled its AI imagery pipeline with the goal of improving prompt understanding and image photorealism. This effort yielded Imagen 3, a next-generation model designed to produce more accurate and varied images in a vast range of styles while reducing risk of bias and exploitation.
As a reaction to their previous failures, the company incorporated stringent content filters that prohibited generation of unsafe or sensitive images. In addition to that, Google set improved guidelines and testing routines to ensure that the AI behaves uniformly in a wide range of situations. The individuals generation currently back in use is controlled by these new safety functions.
Google’s rollout plan is designed to be limited, controlled exposure to the new image generation features. By initially offering the tool to expert and enterprise users, the company is seeking feedback and monitoring system performance before expanding to a broader audience. This slow and cautious rollout is part of Google’s overall goal of building stable AI products.
Read Also : Apple to Reach 40 Billion Dollar iPhone Production in India by FY26
