The system in Honor devices helps with AI fraud detection

Today, we can no longer believe everything we see on the Internet. Creating a fake video in which anyone talks about anything is no longer “atomic science” and thanks to artificial intelligence, even a less skilled person can spread similar material.

The so-called “deepfakes”i.e. artificially created but realistically processed digital media, can imitate people through photos, videos and voice calls in such a way that we can hardly recognize at first glance whether it is a real person or an artificial one.

Fraudsters also use artificial intelligence

This not only poses a security threat, but causes many people don’t trust AI technologies. With the ability to impersonate anyone, cybercriminals can conduct phishing scams or other fraudulent activities with high success. Honor is aware of these risks, which is why it has invested in new technologies to help users of its devices navigate the complex AI environment.

Honor AI Deepfake Detection is a solution implemented directly on the device that detects attempts at fraudulent facial impersonation and immediately alerts users to potential risks. This technology, trained on a large set of videos and images related to network fraud, can quickly detect and warn users about a generated or altered face and protect them from potentially fraudulent activities.

Honor AI recognizes synthetic imperfections

By analyzing spatial features, AI recognizes pixel-level synthetic imperfections, composition boundary artifacts, continuity between images, consistency between the face and hair around the ears, and the placement of synthetic tracks. By examining these subject recognition factors, AI can detect even the smallest anomaly and alert users to it.

You might be interested in:

Honor has implemented protection MagicGuard at the chip level directly into devices as part of the MagicOS system. The security measure consists of three forms of protection which include chipset, operating system and applications. Owners of smartphones and tablets are thus better protected against possible risks. For example, the system alerts them that they are talking to a potential fraudster.