- WhatsApp is reportedly working on implementing a new generative AI feature
- Users may soon be able to generate AI avatars in the popular messaging app
It is the most popular messaging app for millions of users worldwide.
And now, WhatsApp may soon get an AI tool that will give users even more ways to communicate.
Meta’s messaging service is reportedly working on a feature that will allow users to create personalized avatars of themselves in any environment imaginable.
By uploading a collection of photos, WhatsApp users can train Meta’s AI to create realistic digital images that look like them.
However, initial screenshots of the new update look strikingly similar to Apple’s similar ‘Image Playground’ tool, which was announced last month.
WhatsApp is reportedly preparing to roll out a new AI avatar generator feature that will allow users to create images from photos of themselves. Pictured is a leaked screenshot from WABetaInfo
First noticed by WABetaInfoThis update included a new beta version of WhatsApp that could be released as part of a software update later this year.
In a screenshot of the beta feature, users are instructed to “take photos of yourself once, then imagine you’re in any environment, from the forest to space.”
The sample images in the screenshot show a female avatar in different aesthetic styles and backgrounds.
It is believed that users will be able to set them as their WhatsApp profile picture or even send them to friends.
To use this feature, WhatsApp users must upload at least one photo of themselves.
WhatsApp users can access the Meta AI service by typing in a special chat or writing ‘@Meta AI’ in any conversation
These are used to train Llama’s generative AI model, which is responsible for creating the images.
Once the images are uploaded, users simply type “Imagine…” into the in-app Meta AI chatbot and add the description they want to generate.
You can also contact Meta AI via a WhatsApp conversation by typing: ‘@Meta AI imagine me…’.
While users will need to upload an image to access the feature, WABetainfo also reports that the reference image can be removed at any time via the Meta AI settings.
Moreover, the feature is completely optional and users will have to manually activate the AI option to use it.
Currently, the app is still in beta, meaning that development is not yet complete. Meta has not yet provided a date for the full rollout.
Meta AI is available to a small group of users in the US, but is expected to roll out more broadly later this year.
Similar to other AI image generators like Lensa AI or Snapchat’s ‘Dream’ feature, Meta’s new feature appears to use a combination of user-generated images and text prompts.
The most striking similarity, however, is with an AI image-generating feature Apple unveiled at WWDC called Image Playground.
Coming soon as part of Apple Intelligence in iOS 18, the feature will let users create AI avatars using photos of real people.
Apple’s Image Playground lets users create fun images in seconds, choosing from three styles: Animation, Illustration, and Sketch
Image Playground allows users to create images in seconds, choosing from three styles: ‘animation’, ‘illustration’ and ‘sketch’.
A promotional photo appears to show a photo of a person being transformed into a 3D video game-style avatar, dressed in a spacesuit.
It’s just one of several AI features Apple has unveiled, including the option to create AI-generated emojis from text messages.
Users in the UK and EU will have to wait until at least 2025 to access the features, due to an ongoing regulatory dispute between Apple and the EU.
However, it’s not clear whether Apple’s Image Playground will allow these images to be used as profiles or avatars, similar to WhatsApp’s new feature.
Meta declined to comment when contacted by MailOnline.