Thanks to a delivery model capable of working locally, Snap says it can generate augmented reality filters with a simple sentence. The social network could be the first to achieve this technological fantasy.
« Clothing from the 1970s, rue de Rivoli in Paris “. With this simple sentence, Snapchat users will soon be able to generate a tailor-made augmented reality filter. Modification of the face, change of clothing, clipping and change of background… The social network, thanks to a diffusion model, claims to be able to locally generate filters. It could be a small revolution for the photo editing sector, with the introduction of generative artificial intelligence on social networks.
What will be the technical limits of Snapchat’s AI filters?
Unveiled by Snap during the Augmented World Expo (a trade show dedicated to augmented reality in California), the generation of a filter by augmented reality could boost Snapchat’s audiences if it were to be as effective as in the demos published by the business.
All major tech companies dream of a system capable of “teleporting” the user to the location of their choice in one sentence, Snap could be the first to achieve it. Meta, for its applications and virtual reality headsets, also has a great interest in developing such a system. The success of EPIK in 2023, an application that generated a catalog of photos from a few selfies, confirms the public’s interest in this technology.
This content is blocked because you have not accepted cookies and other trackers. This content is provided by Twitter.
To be able to view it, you must accept the use carried out by Twitter with your data which may be used for the following purposes: allowing you to view and share content with social media, promoting the development and improvement of products from Humanoid and its partners, display personalized advertisements to you in relation to your profile and activity, define a personalized advertising profile, measure the performance of advertisements and content on this site and measure the audience of this site (find out more more)
By clicking on “I accept all”, you consent to the aforementioned purposes for all cookies and other trackers placed by Humanoid and its partners.
You can withdraw your consent at any time. For more information, we invite you to read our Cookies Policy.
I accept everything
Manage my choices
In its blog post, Snap says its “Real-Time GenAI” technology is a prototype. Snap says it is capable of generating images directly on the smartphone thanks to lighter and faster models on which its team is actively working. Its goal is to deploy its new AI filters by the end of the year, then implement tools for creators to help them improve their professional filters.
Snap’s announcement nevertheless leaves several uncertainties. For example, will all devices be compatible with these filters which require a lot of computing power? The latest Snapdragon and Apple chips should do just fine, but will the others be completely deprived of AI filters? Snap can always turn to its own servers, but that doesn’t seem to be on the cards.
A few months after the appearance of ChatGPT, Snap was the first social network to add a chatbot powered by a language model (GPT-4 in this case) in its application. Widely used by young people (it overtook Instagram in France in 2022), Snapchat is not afraid to test new things, even if it means getting a long lead over its competitors.
If you liked this article, you will like the following: don’t miss them by subscribing to Numerama on Google News.
