A video appeared to show Ukrainian President Volodymyr Zelenskyy telling Ukrainians to surrender, but it was generated using artificial intelligence. It’s now possible to make pretty much anyone appear to say whatever you want.
- A deepfake is a digitally generated image, video or text that looks real, and almost anyone can create one using existing imagery of a person, Chris Meserole, fellow and director of research at The Brookings Institution, explained.
- Deepfakes are problematic for a variety of reasons, such as being used to easily send out state-sponsored propaganda and being used to send out misinformation during times of crisis, according to Meserole.
- Detecting deepfakes is becoming more difficult as the technology and algorithms being used to detect them are, in turn, being used to generate more deepfakes, he says.
- Democratic governments may look into using deepfakes in the future, but there are some ethical concerns being raised about their use, he explained.
SEE ALSO | Advancing artificial intelligence at NIH
Deepfake concept matching facial movements with a different face of another woman in a photo. Face swapping or impersonation. (Adobe Stock)
Natural Language Understanding - NLU - The Ability of Artificial Intelligence to Understand Human Language - Conceptual Illustration (Adobe Stock)
