https://www.nytimes.com/2023/02/07/technology/artificial-intelligence-training-deepfake.html
Pixabay
Deepfake technology has increased in leaps and bounds, and the potential damage that can be caused by intentional and deceptive videos and images in unimaginable. Deepfakes are videos and images created to misinform. New platforms have been created and are being refined that use artificial intelligence (AI) to spread misinformation through deepfakes.
Governments and other groups can spread misinformation using AI programs that take an image and use it in a video for whatever means the group wishes. Pro and anti-establishment groups have been found to have generated videos to sway public opinion, and the more refined the technology becomes, the more dangerous the spread of misinformation becomes.
Services including Synthesia use AI to generate videos from user prompts. Users select avatars and content that is read by the avatar. The problem is that these avatars may be given official-sounding titles and may appear credible, which poses an issue for vulnerable populations. Such a deepfake could be uploaded onto Facebook or Twitter and if seen by a wide audience, could sway opinions in a damaging way.
Propaganda is being generated using deepfake AI technology. In 2017, a Chinese company called iFlytek created a deepfake video of President Trump that was incredibly convincing. Companies are using tools to produce fake newscasters and authority figures which don’t exist that can be very convincing to those who are unaware of deepfake videos. YouTube has entertaining deepfake videos users have made that add faces onto actors in movie scenes are very convincing. As time goes by, companies including iFlytek and Synthesia will continue to refine their technology, and the AI will make the interface easier and more comprehensive.
Companies that produce deepfake AI technology are responsible in part to ensure their technology and AI protocols do not create dangerous content. Governments need to regulate deepfake technology as well to some degree based upon the message and the potential harm the message could cause. As we move into an age of AI technology advancement, we need to ensure the utility they provide does not create a firestorm of harm.
A Steemit Exclusive