“Ten years ago, deep learning was not on anybody’s radar, and now it’s in everything,” said Pedro Domingos, a computer science professor at the University of Washington.
What Is Deepfake
Deepfake refers to technologies that use AI synthesis technology as the core and generate synthetic algorithms based on deep learning and virtual reality to produce text, images, audio, video, or virtual scenes, mainly including face replacement, voice replacement, and synchronous replacement of both faces and voices.
Since 2018, the update and iteration speed of AI synthesis technology has surpassed all expectations. “Deepfake” technology is no longer limited to technology firms, and the emergence of more and more related applications and online tools allows the general public to experience AI generation in a broader range of scenarios.
AI Face Swap
With tools like DeepFaceLab, AI deepfake technology is also reaching the general public. In addition to the well-known reface app, FaceMagic on mobile and DeepSwap on the website are becoming more popular.
Because of the widespread use of social media and advanced AI synthesis technology, “deep fake” content has the potential to become “Internet hot spots” at any time and from any location. The Tom Cruise face-changing video, which exploded on short videos last year and quickly swept through social media over the world, is strong proof.
Visual effects artist Chris Umé teamed up with Tom Cruise’s top imitators and used deepfakes to create these blockbuster videos. In the video, a “person” who looks and sounds exactly like Tom Cruise is either wearing a floral shirt and performing a “coin trick”, or hanging out in a men’s clothing store on the street, making some huge connections to Hollywood superstars.
Animate Photos
By using the AI face generation service provided by “MyHeritage” created by Israeli company D-ID, users can upload pictures of deceased relatives to generate moving images. The “relatives” in the video can make a series of actions such as blinking, smiling, nodding, etc. Users can obtain a face-to-face visual effect with the “living person”, and use AI synthesis technology to make the deceased “live” in cyberspace.
AI Podcast Editing Software
Descript is a podcast editing software using AI synthetic voice technology. Users can edit or even create their own exclusive audio content through AI voice cloning technology. After using the software’s “overdub” function to clone and generate your own AI vocals, you only need to edit the transcribed text to adjust the audio. Users can directly delete or add text to change the audio content.
When deepfakes technology is gradually “popularized”, Internet users will have more choices and innovations in content creation. In the not-too-distant future, everyone will be able to clone their own face and voice through deepfake synthesis and use them in wider fields such as short videos, live broadcasts, and interactive media.
