- The first time AI deepfakes gained interest was in 2017.
- Alexey Chaplygin is the developer of the DeepFake 2.0.
- The technology can be used for good or evil.
The developer of a new version of “Deepfake” technology is Alexey Chaplygin, who lives in the Netherlands. He spent two years on the project, and has now submitted the first version. In the video, he shows that just one photo is enough to “inject personality” into an arbitrary video.
The neural network first creates a three-dimensional image based on the face or entire body, and then integrates it into the video sequence.
The first time AI deepfakes gained interest was in 2017. It was found on Reddit, with a DNN-based face swapping algorithm. Currently, the term deep fake is used to identify all AI generated videos.
The neural network algorithm for changing the appearance of images and videos has reached a new level — the Deepafake 2.0 system has been created, which allows you to replace not only faces, but also the figure, skin color, and hair.
Its creator believes that the update will allow the technology to be applied in virtual fitting rooms, which will reduce the number of returns to stores and reduce carbon dioxide emissions.
Futhermore, using Deepfake 2.0, Alexey placed Botticelli’s Venus on the frames from “pirates of the Caribbean,” and the face of Anne Frank, who created the most famous diary about atrocities against Jews during the Second World War, on the photo of a girl having a cheerful time.
In addition, as an example of how the algorithm works, Chaplygin “sent” his elderly parents on vacation around the world, which they could never afford. Finally, a separate section of the video is dedicated to demonstrating the use of Deepfake 2.0 to change the video, although the result is still far from ideal.
Chaplygin notes that like all apps, its development can be used for good— for example, for virtual fitting rooms in online stores— or for harm— for stealing someone’s identity just by one photo on Facebook. In his opinion, fake image and video detectors will soon appear in every antivirus software package anyway.
The main goal of Deepfake 2.0 is to “try on” a variety of clothes on your body. Moreover, you can experiment with both the face and all the other basic attributes of appearance. Chaplygin offers all companies or investors interested in developing the technology to finance the project.
He may have done a tremendous job personally, but additional funds are needed for further improvement. Perhaps in the near future, such programs will be installed on smartphones directly by Apple and Android.
AI generated technology will continue to grow. According to Grandview Research, the global artificial intelligence market size was valued at $39.9 billion in 2019, and is expected to grow at a compound annual growth rate (CAGR) of 42.2% from 2020 to 2027.
AI is proven to be the significant revolutionary element of the upcoming digital era. In the digital age of this century, there is going to be a lot of new technologies based on AI in this decade.