The UAE Cybersecurity Council is warning residents to be more selective about which AI tools they allow into their smartphones, particularly when an app or website requests access to cameras and other device features.
The guidance, issued on August 28, comes as generative AI services continue to multiply across app stores, websites and search results. While established AI platforms increasingly integrate cameras, microphones, files and screen-sharing features, those same capabilities can create privacy risks when users hand permissions to unfamiliar or poorly vetted services.
The council says people should verify that an AI service is legitimate and trustworthy before using it, rather than assuming a tool discovered through an online search is safe. Users should also examine requested permissions and avoid granting access that does not appear necessary for the feature they are trying to use.
Camera access is particularly worth scrutinising because modern AI assistants can use live video or photographs to analyse a user’s surroundings. That can be genuinely useful, but it also potentially exposes faces, documents, computer screens, locations and other information visible in the background. The practical rule is straightforward: an AI tool should not receive persistent access to hardware or personal information simply because it asks.
The UAE Cybersecurity Council is also advising residents to keep particularly sensitive information away from AI services. That includes banking and financial details, passports, Emirates IDs, driving licences, medical records and other documents containing personally identifiable information.
This advice matters beyond obviously suspicious AI websites. Generative AI has encouraged users to treat chatbots as general-purpose assistants, meaning people increasingly upload screenshots, documents, photographs and other material without necessarily considering where that information is processed, how long it is retained or what permissions they previously granted.
The warning also fits into the UAE’s broader concern about artificial intelligence becoming part of the cybersecurity threat landscape. Dr Mohamed Al Kuwaiti, head of the UAE Cybersecurity Council, said in June that malicious uses of AI now extend beyond familiar phishing emails into areas including data exfiltration, destructive attacks and more sophisticated cyber operations.
AI-generated deepfakes and misinformation are another part of that equation. More capable image, video and voice-generation systems have lowered the technical barrier to producing convincing impersonations, making verification increasingly important when unexpected requests involve money, credentials or sensitive information.
For UAE residents, the immediate takeaway isn’t to avoid AI tools altogether. It is to treat them like any other application requesting privileged access to a phone: check who operates the service, question why it needs a particular permission, restrict access wherever possible and keep highly sensitive documents out of prompts and uploads.
As AI assistants gain the ability to see, hear and interact with more of our devices, permission screens are becoming a meaningful cybersecurity boundary rather than another button to tap through.
