Dubai Unveils Open-Source AI Tool to Combat Growing Deepfake Security Threats
As organizations worldwide grapple with the escalating sophistication of synthetic media, the Dubai Electronic Security Center has stepped forward with a new technological defense. During the GISEC Global cybersecurity event, officials from the Middle Eastern emirate’s primary regulatory body unveiled an artificial intelligence model designed specifically to identify manipulated video content. Known as Saraab, the tool represents a significant milestone as the first government-backed deepfake detection model developed within the Arab world.
According to statements from the Dubai Electronic Security Center, the regulatory agency intends to make Saraab widely accessible by open-sourcing the software before the end of the year. Officials plan to deposit the model into the Hugging Face library under a standard research and learning license. By granting researchers, security professionals, and technology companies open access to the source code, the regulator hopes to foster collaborative improvement as synthetic video generation technologies continue to evolve rapidly.
Detailing the mechanics behind the software, Dubai Electronic Security Center representatives noted that the model functions by thoroughly evaluating individual faces and overarching video elements to generate an overall authenticity confidence score. Hassan Majab, an official with the regulatory organization, reported that the system achieves an impressive 91% accuracy rate. Rather than providing a simple binary judgment, Saraab delivers a comprehensive descriptive breakdown. The tool assesses every detected face individually and supplies explanatory details outlining the specific markers that indicate digital manipulation.
The urgency surrounding such detection capabilities has been underscored by recent regional incidents. Last year, digital fabrications began circulating widely across social media networks, purporting to show panicked crowds rushing through local transit terminals following missile strikes. Although prominent fact-checking organizations quickly debunked the footage by pointing out visual anomalies—such as automated luggage movement and irregular architectural layouts—the incident highlighted the disruptive potential of unauthorized synthetic media in public spaces.
Chief Executive Officer Yousuf Hamad Al Shaibani emphasized that the introduction of Saraab aligns with regional ambitions to cultivate proprietary technological solutions that can offer protective benefits far beyond local borders. Furthermore, the regulatory body paired the unveiling of the detection tool with two additional foundational compliance initiatives at the GISEC event. These include the Dubai Cyber Skills Framework, established to standardize professional training pathways and role qualifications, and a dedicated auditor certification program designed to measure compliance against local government information security regulations.
As threat actors increasingly incorporate advanced synthetic techniques into their operational playbooks, cybersecurity experts anticipate that collaborative, open-source defense mechanisms will become vital components of enterprise and public-sector resilience. By inviting global peer review and continuous iteration through public repositories, initiatives like Saraab aim to narrow the technological gap between malicious manipulators and institutional defenders.
Source: BankInfoSecurity
