Google Restores Gemini's Ability to Generate Images of People After Addressing Early Issues

Google To Relaunch Image Generation of People on Gemini After Early Issues - News18

Alphabet said on Wednesday it has updated Gemini's AI imagecreation model and would roll out the generation of visuals of people in the coming days, after monthslong pause of the capability. (Reuters) -Alphabet said on Wednesday it has updated Gemini's AI image-creation model and would roll out the generation of visuals of people in the coming days, after months-long pause of the capability. Google had paused its AI tool that creates images of people in February, following inaccuracies in some historical depictions generated by the model. The company said it has worked to improve the product, adhere to "product principles" and simulated situations to find weaknesses. The company was forced to pause its AI image generator tool after its mistakes with historical figures. Many people were unimpressed with the results that Google's Gemini AI image generator offered, which has not only made the company sit up and stop its availability but also admit the mistakes that have been made. Google's Prabhakar Raghavan explained the issue and highlighted the worrying concerns that AI still poses for the world. He said that the AI tool found it hard to differentiate between a wide range of people, and the AI model became extra cautious to avoid any major lapses and be offensive to anyone. "These two things led the model to overcompensate in some cases, and be over-conservative in others, leading to images that were embarrassing and wrong." For instance, AI cannot mix up the characters of German soldiers from the World War and give them a different ethnicity. The fact that Google's parent company is offering the AI image generation feature once again suggests it is satisfied with the results and now happy for the public to give it a shot.

Google thinks Gemini is ready to generate photos of people again

After hitting pause on its artificial intelligence model's ability to generate images of people, Google (GOOGL) is rolling the feature out again. On Wednesday, Google said it is bringing back Gemini's image generation of people in "the coming days," with early access for English-language users of its paid Gemini Advanced, Business, and Enterprise services. "We've worked to make technical improvements to the product, as well as improved evaluation sets, red-teaming exercises, and clear product principles," Dave Citron, senior director for product management at Google, said in a statement. With its latest image generation model, Imagen 3, which powers some of its Gemini Apps, Google has apparently "made significant progress in providing a better user experience when generating images of people," Citron said. The image generator does not allow users to generate "photorealistic, identifiable individuals, depictions of minors, or excessively gory, violent or sexual scenes." In February, Google paused the feature after users pointed out that Gemini was generating historically inaccurate images of people, including racially diverse Nazi-era German soldiers. "Gemini's AI image generation does generate a wide range of people," Google said in a statement at the time. "And that's generally a good thing because people around the world use it. But it's missing the mark here." Google chief executive Sundar Pichai called Gemini's responses "completely unacceptable," and said that the company "got it wrong" in a memo to employees. The tech giant will continue receiving feedback on the image generation tool from early users, and plans to roll it out to more users and languages, Citron said. The company is also rolling out Gems, a feature previewed at its annual I/O developer conference in May, for Gemini Advanced, Business, and Enterprise users. Gems are supposed to enable users to customize Google's AI model to create "personal AI experts on any topic."

Google Gemini AI Update Restores Ability to Generate Images of People

To start, the image generation will be available for Gemini Advanced, Business, and Enterprise users. Google is also rolling out customized versions of its AI, known as Gems. Google is rolling out an update to its Gemini AI that brings back the ability to generate images of people, though it will initially be limited to paying customers. "Over the coming days, we'll...start to roll out the generation of images of people, with an early access version for our Gemini Advanced, Business, and Enterprise users, starting in English," Google says. "We've worked to make technical improvements to the product, as well as improved evaluation sets, red-teaming exercises and clear product principles." Google disabled Gemini's ability to generate any images of people in February after it produced anachronistic historical images, including a Native American man and Indian woman to be representative of an 1820s-era German couple, an African American Founding Father, Asian and indigenous soldiers to be members of the 1929 German military, and diverse representations of a "medieval king of England," among other examples. Gemini Advanced provides access to "the best of Google AI" for $19.99 per month, including access to next-gen models and 1 million token context window. Google says it's "made significant progress in providing a better user experience when generating images of people" with its Imagen 3 text-to-image generator. But unlike the wild images of public figures being produced by xAI's Grok 2, Gemini does not "support the generation of photorealistic, identifiable individuals, depictions of minors or excessively gory, violent or sexual scenes," Google says. The company also warns that "as with any generative AI tool, not every image Gemini creates will be perfect, but we'll continue to listen to feedback from early users as we keep improving." There's no timeline for a wider launch; Google says it will "gradually roll this out, aiming to bring it to more users and languages soon." Imagen 3 is rolling out to Gemini Apps and expanding its availability to users in all languages. Also today, Google released Gems, which are customized versions of its AI. First previewed at I/O in May, Gems do any number of things, from helping you code to editing your writing and even giving advice. Google says it'll be pretty easy to use. Just give Gemini some instructions and a name, and you're off to the races. Gems are launching on desktop and mobile devices for Gemini Advanced, Business, and Enterprise users in more than 150 countries in most languages.

Google Gemini will let you create AI-generated people again

Google is letting its users generate images of people through its Gemini AI chatbot again after pulling the feature earlier this year amid reports of historically inaccurate images, like racially diverse Nazis. In an announcement, Google says it will roll out an early access version of the capability to Gemini Advanced, Business, and Enterprise users in English "over the coming days." This upgrade will be powered by Imagen 3, the latest version of Google's AI text-to-image generator. Google quietly launched Imagen 3 through its AI Test Kitchen earlier this month, and now it's coming to Gemini across all languages. The upgraded tool is capable of generating anything from photorealistic landscapes to textured oil paintings with a description of "just a few words." Google paused Gemini's ability to generate images of people in February after users found it created historically inaccurate images. The upgraded Imagen 3 model comes with built-in safeguards and "performs favorably compared to other image generation models available," Dave Citron, Google's senior director of product management for Gemini, writes in the announcement. This isn't to be confused with Google's new Remagine feature, which lets you incorporate AI elements into the photos shot on the company's new lineup of Pixel 9 phones. Additionally, Gemini won't allow users to create photorealistic images of public figures; content involving minors; or gory, violent, and sexual scenes. "Of course, not every image Gemini creates will be perfect, but we'll continue to listen to feedback from early access Gemini Advanced users as we keep improving," Citron adds. Google plans to expand the ability to create AI-generated people to more users and languages soon.

Google-parent Alphabet to roll out image generation of people on Gemini after pause

Alphabet said on Wednesday it has updated Gemini's AI image-creation model and would roll out the generation of visuals of people in the coming days, after months-long pause of the capability. Google Pixel 9 Pro, Pixel 9 Pro XL, and Pixel 9 launched in India: Price, features, availability Google had paused its AI tool that creates images of people in February, following inaccuracies in some historical depictions generated by the model. The company said it has worked to improve the product, adhere to "product principles" and simulated situations to find weaknesses. Read Comments

Google re-releases Gemini people generator after six-month hiatus: report

Google (NASDAQ:GOOG)(NASDAQ:GOOGL) is preparing to re-release the people generating feature of its artificial intelligence assistant Gemini sometime in the next several days, according to a report by TechCrunch on Wednesday. Google paused the feature back in February after backlash stemming from how the image generator handled race. At the time, Demis Hassabis, head of the research division at Google DeepMind, said the issue would be resolved "in a very short order." The image generator feature will soon be available to users subscribed to paid Gemini plans, including Gemini Advanced, Business or Enterprise, the report said. Google said it has modified Imagen 3, the image-generating model built into Gemini, in order to create "fairness." "Specifically, we evaluate fairness through automated metrics based on the distribution of perceived age, gender, and skin tone in images resulting from generic people-seeking prompts," according to a DeepMind technical report. More on Alphabet, Alphabet, etc. Google: Solid Business, But Regulatory Risks And Potential Recession Google: Attractive Cloud And AI Opportunity Alphabet/Google: Solid Financial Performance And Notable AI Milestones Affordable Android devices, Apple AI, to keep smartphone growth alive: IDC Telegram app CEO is said to be detained in France

Gemini will soon generate AI images of people again with the upgraded Imagen 3

Google's generative AI tools are getting some of the boosts the company previewed at Google I/O. Starting this week, the company is rolling out the next-gen version of its Imagen image generator, which reintroduces the ability to generate AI people (after an embarrassing controversy earlier this year). Google's Gemini chatbot also adds Gems, the company's take on bots with custom instructions, similar to ChatGPT's custom GPTs. Google's Imagen 3 is the upgraded version of its image generator, coming to Gemini. The company says the next-gen AI model "sets a new standard for image quality" and is built with guardrails to avoid overcorrecting for diversity, like the bizarre historical AI images that went viral early this year. "Across a wide range of benchmarks, Imagen 3 performs favorably compared to other image generation models available," Gemini Product Manager Dave Citron wrote in a press release. The tool allows you to guide the image generation with additional prompts if you don't like what it spits out the first time. Citron says Imagen 3 performs "favorably" compared to the competition. It also includes Google's SynthID tool to watermark images, making it clear that they're AI-made and not the genuine article. Citron says the ability to generate people will return in the coming days for paid users, months after Google yanked the feature. He says new guardrails will prevent the generation of "photorealistic, identifiable individuals" -- a far cry from the problematic deepfakes generated by Elon Musk's Grok. Also off-limits are children and (as with other image generators) any gory, violent or sexual scenes. The product manager grounds expectations by saying Gemini's images won't be perfect, but he promises the company will continue to listen to user feedback and refine accordingly. Starting this week, the Imagen 3 model will be available for all users, but reintroducing images featuring people will begin with paid users. English-speaking Gemini Advanced, Business and Enterprise users can expect human image generation to return "over the coming days." Initially previewed at Google I/O 2024, Gems are Google's custom chatbots with user-created instructions. It's essentially Gemini's answer to OpenAI's GPTs, which Google's competitor rolled out late last year. Gems begin rolling out in the next few days. "With Gems, you can create a team of experts to help you think through a challenging project, brainstorm ideas for an upcoming event, or write the perfect caption for a social media post," Citron wrote. "Your Gem can also remember a detailed set of instructions to help you save time on tedious, repetitive or difficult tasks." In addition to the blank slate of custom Gems, Gemini will include premade ones "to help you get started" and inspire new ideas. Prebuilt Gems include: Gems begin rolling out today on desktop and mobile. However, they're only available for Gemini Advanced, Business and Enterprise subscribers, so you'll need a paid plan to check them out.

Google Gemini now lets you create AI-generated images of people -- but there's a catch

The feature was previously available on Gemini, but was disabled in February by Google after users quickly discovered they were able to create disturbing images using the tech giant's AI tool. To prevent exploitative misuse, Google claims that it is implementing strict guardrails. In addition to this news, the search engine giant said that it's rolling out customized Gems, a new feature that was unveiled at Google I/O. Gemini now lets users create AI-generated images of people -- but there are restrictions As mentioned at the outset, Google announced that that Gemini users will now have the opportunity to create AI-generated images of people. As it stands now, if you type in a command that attempts to prompt Gemini to create an AI-generated person, you'll get the following message: "Image generation of people is coming soon to Gemini Advanced. Sign up and get notified when you receive early access at gemini.google/advanced." Well, the wait is now over. Google said that, over the coming days, users will have the opportunity to use Gemini to create AI-generated images of people. Users with Gemini Advanced, Business, or Enterprise accounts will get early access to the feature. Google wants to make clear, however, that this isn't a free-for-all. The search engine claims that it has made "technical improvements" to Gemini ahead of the rollout, as well as "improved evaluation sets, red-teaming exercises and clear product principles." "We don't support the generation of photorealistic, identifiable individuals, depictions of minors or excessively gory, violent or sexual scenes," said Google Gemini's Senior Director of Product Management David Citron. Citron went on to say that Google Gemini is imperfect, but the company will continue to keep an eye on user feedback to improve it. For now, Google Gemini's new people-focused AI-generated feature supports prompts in English, but the company has plans to add more languages soon. Google Gemini runs on Imagen 3. As we reported last week, this Google-developed generative AI model received an upgrade to its capabilities, including improved image quality. Similar to Imagen 2, Imagen 3 uses SynthID, Google's tool for watermarking AI-generated content. A new personalized Gemini is here Google unveiled its plans for customized "Gems" (i.e., personalized variants of Gemini) at Google I/O. Here's how Google described Gems in a mid-May blog post: "You can create any Gem you dream up: a gym buddy, sous chef, coding partner or creative writing guide. They're easy to set up, too." Google claims that you can create a customized Gem by telling it what you want it to do and how you want it to respond. For example, you can use a prompt like, "You're my nutrition coach. Give me a daily nutrition plan. Be optimistic, energetic, and inspiring." And just like that, Google Gemini will create a personalized Gem for you that meets your needs. If you have no idea how you want to customize your Gem, Google said it has created premade Gems for users: Gems are now rolling out to Gemini Advanced, Gemini Business, and Gemini Enterprise users.