AI-Powered Deepfakes Pose Serious Implications for Our Future
Artificial Intelligence (AI) researchers have come up with a terrifying new way to utilize the technology. By using just a single image of someone, researchers are able to create a completely realistic-looking video of them with a trained artificial intelligence system. Referred to as a deepfake, these AI-powered videos are nearly impossible to discern that it is computer generated, and now, they pose a serious threat to our future.
Deepfake videos and audio recordings are now readily found across the web and include some made to emanate Hollywood celebrities, sports stars, politicians, and even Presidents. This has made a lot of people nervous, as it poses serious security implications as well as the potential to spread dangerous propaganda relatively easy.
The spread of these videos and audio recordings has raised concerns as being a potential tool for nefarious parties to spread misinformation, manipulate elections, and even start wars. For example, deep fakes can be created to show a political candidate appearing to say things that could seriously harm their chances for election. They could even be used to broadcast fake emergency warnings for things such as a natural disaster, terrorist attack, or a biothreat. In a more sinister situation, deepfake technology is being used to graft an unsuspecting person’s face into explicit adult movies, making it seem as if that person was in a pornographic film.
Due to their realism, deepfakes can scramble our understanding of truth in multiple ways. They exploit our ability to trust the reliability of evidence we see with our own eyes, turning a completely false narrative into apparent facts. Not only this, they pose a threat to undermine our trust in all videos we see. Knowing deepfakes exist, it can make it difficult to ascertain the genuine truth from falsity.
The quality of deepfakes is only getting better, as AI research becomes more sophisticated. Take for example research coming out Samsung’s AI Research Center. Samsung has figured out how to create a vividly realistic deepfake of anyone by simply stealing your Facebook profile picture. The technology is so innovative, Samsung showed how it can create a deep fake video using a photo of Mona Lisa. The animated clips shows Mona Lisa speaking and moving as if she were real.
Just this week, a deepfake video of House Speaker Nancy Pelosi showed her to be speaking in an impaired manner to President Trump. The video was subsequently shared thousands of times across social media, with countless people believing the video to genuine. Perhaps even more troubling is the fact that this deepfake video is of low-quality, made by someone without extensive knowledge of producing deepfakes. Nevertheless, it passed as being authentic by many who saw it.
While this technology continues to develop, it can be readily accessed and used by anyone with a few hours of their time. Some social media platforms are beginning to implement policies about posting deep fake videos, but for now, there does not seem to be a way to put a stop to using this technology for nefarious purposes.