Meta faces pushback over a new feature that lets users generate AI images using photos from public Instagram profiles without explicit consent from the people pictured. The company says users can opt out, but privacy advocates reject that as insufficient.
The feature, part of Meta's broader push into generative AI tools, allows anyone to feed public profile pictures into the platform's image generation system. Meta claims the opt-out mechanism gives people control, but campaigners argue this inverts standard privacy protections. They want an opt-in system instead, where users must actively consent before their likenesses become training or generation material.
Privacy groups worry the feature creates a backdoor for deepfakes, non-consensual AI-generated imagery, and identity theft. A public Instagram photo, meant for social sharing, becomes raw material for synthetic media. That distinction matters legally and ethically. The burden currently rests on individuals to discover and disable the setting rather than on Meta to request permission upfront.
This move reflects Meta's aggressive AI expansion following its investment in generative models. The company competes with OpenAI, Google, and others in the race to dominate AI tools. But that competition runs headlong into legitimate privacy concerns. Instagram's 2 billion monthly active users represent an enormous dataset for training and deployment.
Meta's opt-out defense has precedent in its past privacy decisions, but regulators and campaigners have grown less tolerant of that approach. The European Union's AI Act and similar emerging frameworks increasingly demand explicit consent for high-risk applications. Using people's likenesses without permission sits squarely in that category.
The outcry signals a broader tension. Tech companies want to move fast with AI features. Regulators and users want guardrails. Meta's current approach satisfies neither camp.