Google has announced that Gemini, its AI tool to rival ChatGPT, now supports AI-generated images of people.
The feature had been available on Gemini for some time but was disabled by Google in February after users quickly discovered they could use the tech giant's AI tool to create disturbing images, with Google claiming it had implemented strict guardrails to prevent abuse.
In addition to this news, the search engine giant said it will roll out customized Gems, a new feature announced at Google I/O.
See: Gemini Live: I used it before you, and it's currently the biggest threat to ChatGPT
Gemini allows users to generate images of people with AI, but it has limitations
As mentioned at the beginning, Google has announced that Gemini users will be able to create AI-generated portraits of people. Currently, when you enter a command to have Gemini create an AI-generated portrait of a person, you will see the following message:
“People image generation will be coming to Gemini Advanced soon. Sign up at gemini.google/advanced to be notified when you receive early access.”
Well, the wait is over. Google says that in the coming days, users will be able to use Gemini to create AI-generated portraits of people. Users with Gemini Advanced, Business, or Enterprise accounts will have early access to the feature.
But Google wants to make it clear that this isn't a laissez-faire approach: The company claims that it has made “technical improvements” to Gemini ahead of the rollout, as well as “improved evaluation sets, red team exercises, and clear product principles.”
Mashable Lightspeed
“We do not support the creation of photorealistic, personally identifiable, or depictions of minors, or excessively gory, violent, or sexual scenes,” said David Citron, senior director of product management at Google Gemini.
Citron added that Google Gemini is not perfect, but the company will continue to pay attention to user feedback to improve it. Currently, Google Gemini's human-centric AI generation feature supports prompts in English, but the company plans to add more languages in the near future.
Google Gemini runs on Imagen 3. As we reported last week, this generative AI model developed by Google has been upgraded with improved image quality and other features. Like Imagen 2, Imagen 3 uses SynthID, a Google tool for watermarking AI-generated content.
Introducing the new personalized Gemini
At Google I/O, Google announced plans for customized “Gems” (i.e. a personalized version of Gemini). In a blog post in mid-May, Google described Gems as follows:
“You can create Gems for gym buddies, sous chefs, coding partners, creative writing guides, or anything you can imagine. And they're easy to set up.”
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 prompts like, “You are my nutrition coach. Tell me your daily nutrition plan. Be optimistic, energetic and inspiring.” Google Gemini will then create a personalized Gem that meets your needs.
If you don't know how to customize a Gem, Google says it creates pre-made Gems for its users.
Learning Coach – Clarifying Complex Topics
Brainstormer — Helps you brainstorm creative ideas and get inspired
Career Guide – Providing advice, recommendations and plans to help you achieve your career goals
Writing Editor — provides constructive feedback on your writing
Coding Partners — Develop your coding skills
Gems is currently rolling out to Gemini Advanced, Gemini Business, and Gemini Enterprise users.
Topic Artificial Intelligence
