THANK YOU FOR SUBSCRIBING
Enterprise Security Magazine | Tuesday, November 29, 2022
FREMONT, CA: Intel (NASDAQ: INTC) claims to have created an AI to determine whether a video has been altered using deep fake technology in real-time. FakeCatcher, a component of the chipmaker's ethical AI research, has a 96 per cent accuracy rate for detecting deepfakes in milliseconds.
The State University of New York at Binghamton's Umur Ciftci and Intel scientist Ilke Demir developed FakeCatcher. They claimed that deep fake videos were now commonplace. Videos of famous people acting or saying things they never actually did or said are probably something they have already seen. Deepfake videos might be challenging for humans to spot. They are increasingly being used to spread false information and deceive consumers, as well as by cybercriminals entering businesses.
Stay ahead of the industry with exclusive feature stories on the top companies, expert insights and the latest news delivered straight to your inbox. Subscribe today.
According to a VMware report from August, 13 per cent more respondents than in a report from 2021 witnessed malicious deep fakes being deployed in cyberattacks. FakeCatcher aims to assist organisations in resolving this issue much more quickly and accurately than current detection approaches, which necessitate uploads and hours to deliver findings.
It runs on a server and connects via a web-based platform using Intel hardware and software. An ensemble of specialised Intel tools on the software creates the optimised FakeCatcher architecture.
While most deep learning-based deepfake detectors now in use hunt for modification in a video's raw data, FakeCatcher searches for hints by examining what makes us human, including minute blood flow signals in a video's pixels. Blood flow signals are gathered across the face, and computers turn these into spatiotemporal maps. Then, via deep learning, they can instantly determine whether a video is authentic or fraudulent, according to the business.
There are many possible applications for this technology. Such detectors might be used by video-centric social media sites like TikTok and Facebook to deter individuals from posting damaging deepfakes. They could also stop media corporations from unintentionally promoting distorted information.
Organisations could potentially use this technology to stop deepfakes from accessing their environments and compromising their systems.
More in News