In summary,
- Google has brought back Imagen 3, an enhanced image-generation tool that can be customized to form an expert team.
- Both functionalities are being released gradually.
According to a recent announcement from Google, Gemini can now produce photos using Imagen 3. The tech giant has now reinstated the feature that lets users create photographs using AI models, after it was discontinued earlier this year. However, as previously mentioned, Google’s Gemini has these capabilities before. Gemini made a U-turn on Imagen after receiving negative feedback on its earlier iteration of the technology, which was used to generate AI images.
Google faced criticism for their image-generation tool back in February. The data points to the tool’s extreme bias and erroneous portrayal of historical events. For instance, inaccurate portrayals of race and more. Following these grievances, Google made the decision to stand aside and address the problems.
And now, after six months, Google is prepared to launch an enhanced version of Imagen 3, an image creation tool. Let’s examine the differences between it.
Gemini’s Imagen 3: A better tool for creating images
Senior Gemini director Dave Citron writes in the Google blog that the business has improved its creative picture generating capabilities. In the upcoming days, the Imagen 3 will be made available to all users. It has enhanced the creative feature and made it more widely accessible to users across all languages.
Google states that “Imagen 3 sets a new standard for image quality, generating images with just a few words.”
Additionally, he said that Imagen 3 has produced “significant progress” in the company’s ability to create human portrayals.
Citron stated, “We’ve worked to make clear product principles, improved evaluation sets, and red-teaming exercises in addition to technical improvements to the product.” Red teaming is the process of testing a product by creating realistic scenarios in order to find any vulnerabilities before it is made available to the general public. There will be flaws in Imagen 3’s creative section, the blog post continued. However, Google is open to receiving user feedback to enhance the tool.
According to Citron, users would not be able to produce “excessively gory, violent or sexual scenes,” recognizable identities, photorealistic imagery, or depictions of kids. He continued, “Google will progressively roll out the feature to more users.”
Create Gems: New Gemini experts
Google has released several new features in addition to Imagen 3. Gemini Advanced, Business, and Enterprise subscribers will be able to create and communicate with Gems—the personalized version of Gemini that was initially shown at I/O—during the next several days. Using this capability, users can tailor Gems to become subject matter experts or hone them for certain purposes.
“You can assemble a team of experts to help you think through a difficult project, generate ideas for an impending event, or compose the ideal caption for a social media post with Gems,” the Google stated. To assist you save time on laborious, repetitive, or challenging jobs, Your Gem may also retain a comprehensive set of instructions.”
Gems are now being made available to Gemini Advanced, Gemini Business, and Gemini Enterprise users in over 150 countries in the majority of languages on desktop and mobile platforms.