WhatsApp integrates Meta AI for in-chat image creation, sparking privacy concerns

WhatsApp is integrating Meta AI to let users generate and modify images directly from their chats. This feature transforms written descriptions into personalized visuals, eliminating the need for external applications. Users can simply open a chat, access Meta AI, and provide a detailed prompt, such as "Create an illustration of a cat astronaut walking on the Moon, video game style." The underlying technology, Meta's Muse Image model, is designed to process complex instructions and text within illustrations.

“

Create an illustration of a cat astronaut walking on the Moon, video game style.

Meta has implemented strict content moderation policies to prevent the creation of explicit material, graphic violence, hate imagery, illegal activities, and deceptive portrayals of real people or public figures. If a prompt violates these rules, the chat will display a notification that the request cannot be processed. The tool also allows users to modify existing photos by applying new styles or altering specific elements through text commands.

“

If an instruction violates these norms, the chat displays a notice indicating that the request cannot be processed.

However, the tool faces criticism for its potential misuse. The Muse Image model can reportedly use @mentions to pull photos from public Instagram profiles for use in creations. This means images of individuals with public profiles could be generated without their knowledge or explicit consent, unless they have specific content reuse protections enabled. Critics, as reported by outlets like Xataka and PYMNTS, warn this could facilitate harassment or identity theft.

“

Muse Image allows the use of @mentions to 'pull' photos from public Instagram profiles and use them in creations.