Experts warn that generating AI caricatures from personal photos can expose users to security risks.

The trend of creating colored caricatures through AI chatbots, based on photographs and personal information, raises alarm bells among cybersecurity experts. Users who upload images of themselves, along with details about their professional role, risk providing scammers with a "treasure" of information. Bob Long, vice president at Daon, emphasizes that these challenges can facilitate the work of criminals by providing a visual representation of users' identities.

The images sent to chatbots are analyzed to extract data about emotions and environment, and this information can be stored and used later. A security breach at companies like OpenAI could allow this data to fall into the hands of criminals, who could use it for convincing impersonations. Experts recommend that users be cautious, avoiding the upload of images that reveal identifying information and checking privacy settings to protect their personal data. OpenAI offers options to exclude data from AI training, but users need to be aware that certain information may still be retained even after deletion.