New AI deepfake technology can now edit videos to change what users are saying. The 3D technology is able to change the speech and mouth pattern to match the desired text.
The software was developed by scientists from Princeton University, Stanford University, The Max Planck Institue for Informatics, and Adobe Research. The technology uses the subject's isolated phonemes (i.e. "ah", "fuh") and then reconfigures them to match the new message. Next, the software combines phonemes, visemes, and 3D facial models to create new visual footage. The edited videos were perceived as real by 60% of participants surveyed. This new technology has a number of implications. It may be able to help Hollywood translate movies.
Image Credit: SIGGRAPH 2019
What Makes This Trend Stand Out
- AI Deepfake Technology
- Opportunity to disrupt the media and entertainment industry by creating realistic video content with altered speech.
- Speech Editing Software
- Potential to revolutionize the advertising industry by allowing advertisers to change speech in videos to deliver targeted messages.
- Realistic Video Alteration
- Disruptive innovation opportunity in the security industry through the creation of convincing but false evidence in video footage.
Sectors Adopting This
- Media and Entertainment
- The AI deepfake technology can revolutionize the creation of video content with altered speech, offering new possibilities for the industry.
- Advertising
- Speech editing software can provide advertisers the ability to modify videos and deliver tailored messages to specific target audiences.
- Security
- The realistic video alteration capability poses a disruptive innovation opportunity to the security industry by enabling the creation of convincing, but false, video evidence.
