Artificial intelligence technologies, which develop every day, continue to surprise us. “Animate Anyone”, which has been on the agenda on social media lately, is one of them. This model, which will cause us to be afraid when sharing photos on the internet, was developed by the artificial intelligence team of Chinese technology giant Alibaba.
To put it briefly, Animate Anyone is a model that can convert photos into videos. As if deepfake photos were not enough, this artificial intelligence tool, which can now cause us to see fake videos, literally brings an era to photo-to-video conversion technologies.
Animate Anyone has the ability to create a video with a single photo of you
Animate Anyone can create several-second videos from a single photo of any person. In the shared images, we can see that the videos created by the model are eerily realistic.
Videos can be created not only from human photographs but also from photographs such as anime characters.
https://twitter.com/twitter/status/1730876744188534865
*If you cannot access the post, you can click the link here.
The model, described in a study published in arXiv, can convert not only human images but also different images such as cartoon and anime characters into video. Developers at Alibaba have also opened a Github page for Animate Anyone.
The model, which has some flaws, is not yet ready for wide release
https://twitter.com/twitter/status/1730876691755450572
*If you cannot access the post, you can click the link here.
When we examine the videos, we see that the people in the photographs do not experience major deterioration. In one video, the anime character comes to life, and in the other, Lionel Messi starts to move. Of course, since it produces short videos, it can even be used to create fake TikTok videos.
But Animate Anyone isn’t perfect, as you might expect. Just like other productive models, she has problems with hands and eyes. Or it may have difficulty when necessary to perform actions such as rotating in videos, resulting in some strange images. The developers already underline that it is not completely ready yet. Still, we can say that it is a significant leap forward for such technologies.
According to the Alibaba team, Animate Anyone is not ready to be shared with the world. “We are actively working to prepare the demo and code for public release,” the team said in a statement on Github. “We cannot give an exact release date for now.” says.
We should not forget the risks that such technology will bring.
https://twitter.com/twitter/status/1730876833216827490
*If you cannot access the post, you can click the link here.
It would not be wrong to say that this type of technology will cause some concerns. While deepfake photos are already causing big problems, it will not lead to good results if fake videos start circulating on the internet.
The possibility that a malicious person can make that person do whatever he wants with just a single photo is very frightening. Hopefully, a solution will be found when this technology becomes widely available. However, it is certain that social media will be flooded with fake TikTok videos created by artificial intelligence much sooner than we expect.
24037
