100% of the cartoons created by ChatGPT may include hidden technical metadata in the image

To obtain a caricature, the user initiates a conversation with ChatGPT and requests an image based on their message history. In many cases, the assistant requires a photograph to refine the visual details

09 of february of 2026 at 15:35h
100% of the cartoons created by ChatGPT may include hidden technical metadata in the image
100% of the cartoons created by ChatGPT may include hidden technical metadata in the image

The request for personalized cartoons through ChatGPT has spread in recent days among users from Barcelona, Girona, and Tarragona, especially in neighborhoods where social media activity is intense. The Instagram and X platforms have seen an increase in illustrations generated by artificial intelligence that show physical features, work references, and personal details of those who request them.

The process and the information exposed

To obtain a caricature, the user initiates a conversation with ChatGPT and requests an image based on their message history. In many cases, the assistant requires a photograph to refine the visual details. This procedure, apparently harmless, involves the delivery of personal data that may include names, occupations, tastes, and habits

Cybersecurity experts have warned that the publication of these images can facilitate the identification of a person even if explicit data such as address or phone number are not shared. Generated and shared images can contain hidden metadata, such as the location, date and time the photo was taken, the device used, and other technical details.

Privacy Policy and Associated Risks

According to OpenAI's privacy policy, all information provided by users may be temporarily stored and used for the training of AI models. The company reserves the right to use this data to improve its services, develop new features, or for commercial purposes. The exact amount of time data remains on the servers is unknown, although OpenAI assures that it is not stored indefinitelyThe risk increases if the images include minors, vulnerable people, or information that could be used for identity theft. In case of a security breach, the photos and their metadata could be exposed and used for phishing attacks, digital harassment, or social engineering campaigns

Cybersecurity Experts' Recommendations

  • Avoid uploading sensitive images to artificial intelligence platforms.
  • Remove metadata before sharing photographs.
  • Limit the amount of personal information provided in requests.
  • Review the privacy policies of each platform before sharing any file.

The trend of personalized cartoons continues to grow in Catalonia, but specialists insist on the importance of taking extreme precautions to protect users' privacy and digital security