Home / Technology / Meta's AI Lets Anyone Deepfake Your Likeness
Meta's AI Lets Anyone Deepfake Your Likeness
8 Jul
Summary
- New AI model allows users to create images of others using their likeness.
- Public Instagram accounts are at risk of deepfake creation by anyone.
- Users can adjust settings to prevent their likeness from being used in AI.
Meta has launched Muse Image, a new AI model designed for generating and editing images, now accessible through the Meta AI app, Instagram, and WhatsApp. This advanced model can handle complex requests and create composite photos, posing a significant deepfake risk for individuals with public Instagram accounts.
Users with public accounts over 18 can have their likeness incorporated into AI-generated images if their username is included in prompts. This capability was demonstrated by creating a deepfake of a CNET colleague in under a minute. Meta offers a privacy setting within Instagram to prevent others from reusing content and likenesses for AI features.
Users can opt to disable the reuse of their content for AI purposes via Instagram Settings. Additionally, the Meta AI app offers settings under 'Your Likeness' to control who can use your image. While Meta states its models have safeguards against illegal content, the potential for misuse remains a concern, echoing issues seen with previous AI tools.
