UNAIR NEWS – Lately, there has been a surge in the photo or video manipulation using Artificial Intelligence that harms individuals on social media. In ‘X’, several users have claimed to be victims of photo manipulation, where images of them fully clothed were transformed into explicit content. Manipulating images or videos with AI to create new content that appears authentic and convincing is known as ‘Deepfake’. How does it work, and what are the threats?”
Aziz Fajar SKom MKom, lecturer of the Faculty of Advanced Technology and Multidiscipline Universitas Airlangga, explained that deepfake is one of the applications of AI models commonly used to alter pixels in images. “By changing the pixel values in an image, the modified one will differ from the original,” he said.
To UNAIR NEWS, the Data Science lecturer conveyed that the AI application is capable of altering facial appearances, making it susceptible to misuse.
“Deepfake can be employed to change someone’s face, allowing the creation of misleading videos or images. In reality, the victim may have never engaged in such activities,” he explained.
How to spot deepfake?
The phenomenon of AI-generated fake images and videos raises public concerns, especially for those active on social media. Often, the victims experience stress due to tarnished reputations in their communities. Moreover, there are still many people who consume information from unreliable sources.
Aziz mentioned that there are now applications capable of detecting deepfakes. One of them is Microsoft’s Video Authenticator Tools. This Microsoft-developed application can assist users in spotting fake photos and videos.
“This software can detect AI-generated fake images and videos,” he said.
“Nevertheless,” he added. “The software may still fail to detect it.”
The Machine Learning lecturer emphasized the need for daily advancements in AI detectors since deepfake technology is expected to continuously evolve. Not only that, but the value of content in society will vary.
“Therefore, it depends on media consumers being willing and able to think critically, investigate sources, or simply believe outright,” he concluded.
Author: Syifa Rahmadina
Editor: Feri Fenoria
Read also
OpenAI introduces Sora, text-to-video tool, here’s what a UNAIR AI analyst says
Revealing impact of Apple Vision Pro on tech competitive market
