Google Rolls Out Enhanced AI Image Generator

Google has launched the latest version of Imagen 3, its AI-powered text-to-image generator, for users in the United States, according to a report from VentureBeat. Available through Google’s AI Test Kitchen, Imagen 3 is designed to produce images with enhanced detail, richer lighting, and fewer distracting artifacts compared to earlier models.

The company initially introduced Imagen 3 during its I/O event in May, but it seems the tool has only recently become widely accessible via the Vertex AI platform. Some Reddit users began testing Imagen 3 last week, and Google released a research paper on the technology this Tuesday.

Similar to other AI image generators, Imagen 3 allows users to create detailed images from text prompts. Additionally, it offers the ability to edit images by selecting a specific area and describing the desired changes.

There are some safeguards in place with Imagen 3, as the tool refuses to generate images of public figures like Taylor Swift and will not produce images of weapons. Although it avoids generating named copyrighted characters, you can still create similar images by describing the character.

For instance, I managed to generate images that closely resemble Sonic the Hedgehog and Mario, while a colleague was able to create characters that look like Mickey Mouse. The tool also generated logos for companies like Apple, Macy’s, Hershey’s, and even Google, as shown in the image at the top of this article.

Despite these relatively lenient safeguards, Imagen 3 is notably different from Grok, the AI image generator on Elon Musk’s X platform, which has been used to create controversial content, including images involving drugs, violence, and public figures in compromising situations.

However, Google’s AI tools haven’t been without their own issues. Earlier this year, Google disabled image generation in its Gemini AI chatbot after users discovered it was producing historically inaccurate images.