Scientists Develop Advanced Tool for Detecting Fake Content
Scientists have introduced a new tool for detecting fake content based on artificial intelligence capable of detecting manipulations at a subpixel level. This technology, known as MISLnet, provides extremely high accuracy in identifying artificially created images and videos.
New Approach Outperforms Traditional Methods
Using MISLnet algorithms, scientists were able to outperform traditional methods of detecting fakes, which are no longer effective in detecting manipulations in artificially created videos. The new approach allows for identifying algorithmic traces left by neural networks during file processing.
High Speed and Accuracy in Identifying Changes
The MISLnet technology was developed based on long-term research and can analyze all changes in images, including addition, movement, and deletion of elements, done at lightning speed. It is trained to distinguish differences in pixel values between generated AI content and original photos and videos.
Unprecedented Accuracy
The current accuracy of the MISLnet tool is an incredible 98%, significantly outperforming other fake content detection systems, which usually only achieve 93%.
A Powerful Tool Against Misinformation
The new MISLnet technology is a significant step forward in combating misinformation and manipulation in the media space, providing a reliable tool for detecting artificially created content.
