UAE: Used AI to create cartoon avatars? You may be risking data theft, deepfakes

Khaleej Times

6 days ago

Have you uploaded your photo to an AI app to get a cute cartoon version of yourself? You might want to think twice — an authority in Abu Dhabi issued a warning on Monday about serious privacy risks that go beyond the fun.
The Department of Government Enablement-Abu Dhabi has alerted the publich of concerns around the growing popularity of AI-powered applications that transform personal photos into cartoon characters or digital avatars, cautioning that many users are unknowingly exposing sensitive biometric data.
In a statement posted through X, the authority highlighted that while uploading such images may seem harmless, it can pose serious privacy and security risks, as this data could be misused for commercial exploitation or malicious purposes, including identity theft and fraud.
Stay up to date with the latest news. Follow KT on WhatsApp Channels.
The department also shared a video addressing the risks associated with these applications. By using just one photo, users have trained AI to recognise and use their face.
The video further warned about potential misuses of personal images, including the creation of fake accounts, fraudulent schemes, and deepfake videos emphasising that users are not just sharing a photo but giving away their biometric identity.
3 tips
To help users protect their privacy, the Department provided several recommendations:
Delete unnecessary apps and photos you previously uploaded on AI apps
Limit app permissions, and block full access
Share knowledge among your family and friends to raise awareness on the underlying risks
Earlier this year, cybersecurity experts raised similar concerns about the viral Ghibli trend that transformed personal photos into Studio Ghibli-inspired avatars.
Nicolai Solling, Chief Technology Officer at Help AG, warned that when creating such avatars, users often upload photos that capture detailed facial features, which can include biometric data that is irreplaceable if compromised.
'The use of these avatars introduces significant risks, primarily related to personal data,' Solling had said. He pointed out that uploaded images may contain hidden metadata, such as device details and location, which could inadvertently expose sensitive information about the user's whereabouts.
Many AI-driven platforms utilise user-uploaded data to enhance their algorithms, and unless users are paying for the service, their images, including facial data, may be used to train AI models, raising serious concerns about data ownership and privacy.
Social media experts also highlighted the surge in user engagement driven by AI-generated avatars, emphasising that while these avatars allow for creativity and self-expression, users must remain vigilant about the potential privacy implications.