Defudger

Defudger is developing a three layer detection system that is able to verify the authenticity of digital content. Our solution reveals forgery and falsification of images and videos with high accuracy. Images and videos that were taken out of context and reused with modified metadata, so called shallow fakes have been around for some time, however the advancement of AI and wide availability of CPU/GPU power has enabled bad actors to easily forge videos, that are nearly indistinguishable from real ones and can be used for malicious and unethical purposes. Videos have been regarded as proof that something has happened; however, today anyone can swap faces of politicians or other people just by following a step-by-step guide available online. With time, this will only get easier and videos like these can spread like wildfire on social media, just like fake news in 2016, but now on steroids, in the form of fake videos. This could possibly happen already in 2020, just in time for the next US elections. Our solution allows media outlets (online, TV, news agencies, journalists) and platforms (social media sites, video sharing platforms and websites) to check the authenticity of images and videos, preventing misinformation of the public and all unwanted consequences related to forged content. The uniqueness of our solution relies on bundling different state-of-the-art technologies: forensic image analysis, metadata validation, reverse image search, AI, machine learning, digital watermarking and blockchain verification.

show less