Skip to content
Newsroom

Explore coverage of SANDS Lab’s technology, products, business performance, and key developments—all in one place. We’ve gathered major stories about SANDS Lab from media outlets in Korea and around the world.

SANDS Lab unveils deepfake-detection AI agent at ACDC 2025

Dec 3, 2025Press Coverage

SANDS Lab showcased a deepfake detection and analysis AI agent at the '2025 ACDC' AI Cyber Defense Contest.

AI security firm SANDS Lab (KOSDAQ 411080, CEO Kim Ki-hong) participated in the '2025 ACDC (AI Cyber Defense Contest)' held December 1-2 at COEX in Seoul, showcasing an AI agent built on deepfake detection and analysis technology.

Hosted by the Ministry of Science and ICT and co-organized by the Korea Internet & Security Agency (KISA) and the Korea Information Security Industry Association (KISIA), ACDC 2025 is a public hands-on exhibition themed 'AI for All, Security for All,' featuring an AI hacking-defense contest and AI security insight seminars among other programs.

At the Safe AI World pavilion, SANDS Lab ran a hands-on program letting visitors generate and detect deepfake images themselves, and demonstrated the technology going into its deepfake-detection service 'Fakecheck 3.0' along with an AI agent model still in development.

While Fakecheck 2.0 focused on detecting images created by generative AI, the technology SANDS Lab has now advanced targets adversarial images — deepfakes with carefully engineered noise designed specifically to evade detection. Fakecheck 3.0, now in development, aims to analyze every type of deepfake including such adversarial attacks, and will also incorporate SANDS Lab's own AI agent.

On the first day, Minister of Science and ICT Bae Kyung-hoon visited the Safe AI World pavilion and personally experienced SANDS Lab's deepfake-detection AI agent model, with CEO Kim Ki-hong leading the demonstration and introducing the company's data-driven AI applications and its AI agent development in cybersecurity.

On day two, a 'cybersecurity AI dataset excellent utilization results sharing session' was held concurrently. SANDS Lab, which has carried out KISA's cybersecurity AI dataset project every year, presented its key results under the theme 'Building and utilizing an APT-based threat profiling AI dataset.'

CEO Kim Ki-hong said, "In the agentic AI era, cybersecurity AI agents will play an ever more important role across the full detection-judgment-response process. The technologies SANDS Lab showed at ACDC 2025 will continue to be refined as core elements of AI-agent-based security going forward."

NextList