The latest AI boom is being abused in various forms, with Deepfake porn becoming the most popular of them all. Here are some tools that are trying hard to solve this plague.
The number of deepfake porn videos, images, and audio content is growing like wildfire. With the boom of generative AI, deepfake technology has become one of the most abused aspects of AI worldwide. People from all backgrounds are prone to become victims of this easy-to-access technology. However, there are new tools and preventive methods to help netizens find a safer way to navigate online.
Deepfake porn is a fake video, image, or audio of a real person, mostly generated without their consent. It can create an original output or alter existing content to place someone else’s face on it. The art of mimicking with finesse has made deepfake videos highly attractive, which helps the abusers earn quick loads of huge money.
The Underlying Problem
The simplest solution to this would be to ban deepfake video technology completely. But, the potential benefits of deepfake videos make it tough to eliminate from technology evolution perspective. That’s why, the deepfake porn problem is one of the toughest to mitigate, let alone eradicate. A study revealed that nearly 96% of deepfake videos are of adult content. While the content is evidently fake, the traum emerging out of it is pretty real.
The underlying issue of deepfake technology is that the videos, images, or audios are almost impossible to trace. Even though most creators end up generating below average quality content that clearly indicates AI generative outputs, the public attention does not really care about it. Deepfake porn videos have become the top streamed videos globally, thanks to the lack of empathy in the general public. Lack of regulations also create a favorable environment for digital deepfake creators to remain undetected. Ease of access and usage of deepfake AI tools is adding more fuel to the deepfake content production.
Probable Solutions
Multiple tools have emerged to identify deepfake videos created by AI. Tech giants like Google and Meta are also rapidly developing AI capabilities that identify and declare AI-generate content. Digital watermarks are being proposed as part of the regulation process for AI-generate content. ChatGPT is working on creating traces of the content’s origin in line with the Coalition for Content Provenance and Authenticity (C2PA) standards.
Dedicated AI-powered platforms are available for free to help users find the source and remove the content from the creators’ end.
Poison Pills are defensive tools designed to blurr the image whenever it detects nudity in it. Another application, Nightshade, created by the researchers at the University of Chiago, aims to mess up the images everytime they are uploaded on an AI platform. The tool will show the images correctly only when it is intended for humans. Atleast 10 states have implemented a patchwork of legal protection approaches for deepfake victims. Indian lawmakers fasttracked their regulatory functions after an Indian celebrity’s deepfake video went viral, leaving crores of citizens in a fright. The possibility to switch off the deepfake creation capability of AI tools is nearly zero. But, they can be brought into strict regulation and controlled with adequate technology in place. How soon, we don’t really know.
