Dubai is making significant strides in cybersecurity with the introduction of SARAAB, an advanced open-source AI model designed to detect deepfakes. This innovative solution, developed by the Dubai Electronic Security Centre (DESC), leverages cutting-edge technology to analyze facial characteristics and movements in videos. By identifying subtle synthetic alterations, SARAAB aims to enhance the integrity of digital media in an age where misinformation can easily spread.

Overview of SARAAB

Anticipated to be launched by the end of 2026, SARAAB was first revealed at the GISEC Global 2026 event, held from September 16 to 18 at the Dubai Exhibition Centre in Expo City. This initiative, which is a product of a dedicated Emirati team within DESC, represents a proactive approach to managing the risks associated with deepfake technology. According to reports, SARAAB is not only expected to function effectively in identifying manipulated video content but also to offer a reliable resource for organizations that need to verify video authenticity.

Mechanisms Behind Deepfake Detection

SARAAB employs a unique methodology to assess video integrity by utilizing a heat map to scrutinize facial features for hidden synthetic layers. As explained by Hassan Majid Alkhazraji, a Senior Artificial Intelligence Executive at DESC, the model can thoroughly analyze an entire two-minute video, even if the digital manipulation is restricted to just the last 30 seconds. This capability is especially crucial in today’s media landscape, where disinformation can occur in moments. The DESC claims that SARAAB achieves an accuracy rate of 91% in detection. However, it is important to note that the methodology for this performance metric has not been independently verified, leaving questions about its reliability in real-world scenarios.

The Open-Source Approach and Its Importance

One of the standout features of SARAAB is its open-source nature, which DESC plans to extend to developer platforms. This initiative will allow artificial intelligence engineers, cybersecurity professionals, and researchers to engage with the model, further testing and refining its capabilities. While Hugging Face is mentioned as a prospective platform for launching SARAAB, no specific timelines or public model pages have yet been made available.

The decision to release SARAAB as an open-source tool can vastly benefit organizations aiming to verify dubious video content. However, it is critical to interpret the results within context. A detection score from an AI model does not definitively prove that a video is either authentic or manipulated, especially given the rapid evolution of synthetic video technologies.

Additionally, DESC’s announcement included the unveiling of other initiatives at GISEC Global 2026, such as the Dubai Cyber Skills Framework and an ISR Auditor Certification Programme. These developments emphasize a broader commitment to building effective local cybersecurity tools and refining the skills necessary to utilize them effectively.

Dubai’s dedication to advancing its deepfake verification capabilities aligns with the UAE Cybersecurity Council’s efforts to underscore the pressing security challenges posed by synthetic media. As the use of deepfakes continues to escalate, the SARAAB model is poised to become an essential component of the region’s digital media verification toolkit.

In conclusion, the introduction of SARAAB highlights Dubai’s forward-thinking approach to cybersecurity. By harnessing the power of open-source AI technology for deepfake detection, the city stands at the forefront of safeguarding digital authenticity, creating a framework that supports continuous innovation and collaboration among industry experts.