Dubai, September 16, 2026: Dubai is preparing to launch a new artificial-intelligence system designed to detect deepfake videos, marking another step in the emirate’s efforts to strengthen cybersecurity and protect people and organisations from increasingly sophisticated AI-generated content.
The system, called SARAAB, has been announced by the Dubai Electronic Security Centre (DESC) during GISEC Global, the major cybersecurity exhibition and conference taking place at the Dubai Exhibition Centre in Expo City Dubai from September 16 to September 18. According to DESC, the model is expected to be launched by the end of 2026 and has achieved an accuracy rate of 91 percent during testing.
What exactly is a deepfake?
A deepfake is digitally manipulated or AI-generated content that can make a person appear to say or do something that never actually happened. Modern AI tools can modify faces, expressions and voices to produce videos that can look remarkably realistic.
This technology can have legitimate entertainment and creative applications, but it can also create serious cybersecurity problems. Fake videos can potentially be used for fraud, impersonation, misinformation, blackmail, identity-related scams and social engineering.
Dubai Police has previously warned the public that some images and videos may appear authentic while actually being generated or manipulated using artificial intelligence. Its advice has been straightforward: people should verify the source before trusting or sharing suspicious content.
How will Dubai’s new system work?
The new system, named SARAAB, is designed to analyse videos and identify portions that have been manipulated.
One interesting feature is its ability to examine a video from beginning to end rather than simply giving the entire video a single classification. For example, if someone uploads a two-minute video where only the final 30 seconds have been manipulated, the system can analyse the entire recording and identify the manipulated section.
The system also uses a heat map to highlight facial areas where the AI believes manipulation has occurred. This could help investigators understand not only that a video may be fake, but also where the manipulation is located.
That distinction is important because deepfake videos are becoming increasingly sophisticated. A fake video may contain mostly genuine footage with only a person’s face, mouth movements or other facial characteristics being digitally altered.
Why is this important for Dubai?
Dubai has positioned itself as a major technology and digital-economy hub, which means cybersecurity has become an increasingly important part of its development strategy.
Artificial intelligence is being adopted across government services, financial services, transportation, healthcare, real estate and other sectors. At the same time, the same technology can be used by criminals to create more convincing scams.
Dubai’s Electronic Security Centre has already been working on policies and initiatives focused on securing AI technologies. In 2024, DESC launched the Dubai AI Security Policy, which was designed to encourage secure AI adoption while addressing electronic-security risks.
The new deepfake-detection model represents a more practical application of that broader cybersecurity approach: rather than only creating rules for AI, authorities are also developing technological tools to identify AI-generated threats.
Could this help fight scams?
Potentially, yes.
Deepfakes can be particularly dangerous when combined with other forms of cybercrime. For example, criminals could theoretically create a fake video or audio message appearing to come from a company executive, business owner or another trusted individual and use it to convince someone to transfer money or reveal confidential information.
Similar techniques can also be used in fake video calls, identity impersonation and phishing campaigns.
However, a 91-percent accuracy rate should not be interpreted as meaning that every deepfake will automatically be detected. No detection system is perfect, particularly as generative-AI technology continues to evolve. The technology will therefore likely need continuous testing and updating as new methods of creating synthetic media appear.
What does it mean for ordinary people?
For residents in Dubai, the development is another reminder that seeing is no longer necessarily believing online.
If you receive a surprising video involving a public figure, company representative, friend or family member, it is worth checking the original source before forwarding it. Suspicious requests for money, passwords, OTPs or personal information should receive additional verification through an independent communication channel.
The UAE has already been increasing public awareness around AI-generated content and digital fraud. Dubai Police, for example, has advised residents to verify the source of suspicious images and videos rather than relying only on how realistic they appear.
The bigger picture
Dubai’s SARAAB project reflects a broader technological race: AI is being used to create increasingly realistic synthetic media, while AI is simultaneously being developed to detect it.
The effectiveness of such systems will depend on how quickly they can adapt to new deepfake-generation techniques. For Dubai, the launch of SARAAB adds another layer to its cybersecurity infrastructure and demonstrates the growing importance of AI security as digital content becomes harder for humans to authenticate visually.
In short: Dubai is preparing to launch SARAAB, an AI-powered deepfake detection model with a reported 91% accuracy rate, by the end of 2026. The system will analyse videos and use heat maps to identify facial areas that may have been manipulated, potentially helping authorities and investigators deal with AI-generated fraud and misinformation.