SANDS Lab launches a deepfake detection service available to everyone
Free, easy and convenient access for everyone as a contribution to society Opened to the public to help eradicate digital crime and ease public concern Security industry’s first deepfake detection service offers a practical response
SANDS Lab launches a deepfake detection service available to everyone
- Free, easy and convenient access for everyone as a contribution to society
- Opened to the public to help eradicate digital crime and ease public concern
- Security industry’s first deepfake detection service offers a practical response
Photo 1. FakeCheck deepfake detection service homepage (Source: SANDS Lab)
AI-based cyber threat intelligence (CTI) security specialist SANDS Lab (KOSDAQ 411080, CEO Kim Ki-hong) announced on the 5th the launch of FakeCheck (fakecheck.io), a service that detects deepfakes, one of the harmful uses of generative AI.
Users can visit FakeCheck’s website and upload a suspected deepfake image to receive an authenticity assessment within seconds. It focuses on its core function, removing cumbersome registration and authentication so that people of all ages can use it easily.
The service is free and available to everyone. To help address a social issue facing the public, SANDS Lab has built it in-house and is covering all maintenance costs, including cloud services, as a social contribution. It believes preventing secondary and tertiary harm from deepfakes takes priority over immediate service revenue.
SANDS Lab was already conducting a national project to build a public-facing platform with functions like FakeCheck through R&D on detecting and responding to harmful uses of generative AI. Less than six months after work began, those harms intensified into a social issue. The company therefore opened the service to the public as part of its responsibility to help eliminate digital crime and alleviate public anxiety.
Photo 2. FakeCheck deepfake detection service (Source: SANDS Lab)
FakeCheck uses multiple AI models to identify manipulated human face images. Because it is designed for crime prevention, it does not assess deepfake images of animals or objects. A representative explained that it was released as an open beta rather than a fully completed service, so continued feedback and updates may expand its targets and improve detection rates, with video manipulation detection potentially added later.
CEO Kim Ki-hong said, “FakeCheck is a practical response created by our employees working together to meet society’s need for countermeasures against deepfake incidents. This launch strengthens the foundation for the public-facing platform to curb harmful uses of generative AI that we are developing under the Ministry of Science and ICT and IITP’s core information security technology program. We will lead efforts to contribute to society through AI technology for future cybersecurity.”
Photo 3. Actual deepfake image analysis results (Source: SANDS Lab)