Meta Must Put Consent Before AI Data Collection
Mark Zuckerberg
Chief Executive Officer
Meta Platforms
Meta Must Put Consent Before AI Data Collection
Dear Mr. Zuckerberg,
We, the undersigned organizations, write to urge Meta to commit to a simple principle that will be incredibly important in the Age of AI: people should affirmatively choose whether their identity, likeness, voice, and personal experiences may be used to power artificial intelligence systems.
Although Meta recently withdrew its Muse AI image generation feature following widespread public backlash, the underlying concerns remain unresolved. Before being pulled from the market, Muse permitted users to generate AI-created images depicting individuals with public Instagram accounts through a default opt-out system rather than obtaining affirmative consent. Consumers were automatically enrolled without their permission, generally received no notice when AI-generated images involving their likeness were created, and were required to navigate complex privacy settings to prevent future use.
The widespread pushback, and subsequent market removal of Muse, served as an important acknowledgement that the public expects stronger privacy protections. But it does not resolve the broader direction Meta appears to be taking. Public reporting indicates that Meta is developing increasingly sophisticated AI-powered wearable technologies capable of continuously capturing images, audio, and other information about users and the people around them. These systems reportedly create searchable records of everyday life, extract detailed metadata about individuals’ activities and relationships, and may ultimately be used to improve Meta’s AI models. Reports also suggest that future devices may reduce or eliminate visible indicators that recording is occurring, making it increasingly difficult for bystanders to know when they are being observed.
This is an upside down approach to privacy. Rather than asking permission before collecting and repurposing deeply personal information, Muse and new technologies reportedly under consideration by Meta place the burden on consumers to discover, understand, and disable complicated privacy settings after the fact. Even more troubling, bystanders often have no meaningful opportunity to consent at all rendering any sense of privacy obsolete.
People do not surrender control over their identity simply because they maintain a public social media profile or exist in public spaces. No person should have to fear that their biometric information will be captured by a technology company simply because they exist in public. Consumers reasonably expect that sharing photographs with friends, walking through a neighborhood, attending school, shopping, or participating in community life does not constitute blanket permission for those experiences to become commercial training data for AI systems.
Privacy is not merely a preference setting, it is a fundamental component of human dignity, individual autonomy, and civil rights and civil liberties.
Accordingly, we urge Meta to:
- Obtain genuine, clear, affirmative consent before collecting or processing biometric information, including to identify a person
- Cease business and data practices that replicate a person’s voice or likeness, or train AI models
- Ensure that privacy protections are easy to understand, easy to exercise, and enabled by default where appropriate.
- Publicly commit that meaningful affirmative consent, not data maximization, will guide Meta’s responsible innovation principles.
Meta can demonstrate that technological leadership includes respecting the dignity, autonomy, and informed choices of the people whose lives make these technologies possible. We urge Meta to adopt these principles before introducing additional AI products that further erode personal privacy.
We look forward to your timely reply.
Public Citizen
American Civil Liberties Union
Autistic Women and Nonbinary Network
Center for AI and Digital Policy
Center for Oil and Gas Organizing
Climate Crisis Working Group of Moore Co.
Color of Change
Common Cause
Common Sense Media
Consumer Action
Consumer Federation of America
Demand Progress Education Fund
Distributed AI Research Institute (DAIR)
Encode AI
Evitable
Innovate EDU
Innovation for Everyone
MIT Data + Feminism Lab
MPower Change
National Black Justice Collective
NoSo Connection Collective
Open MIC (Open Media and Information Companies Initiative)
Oregon Consumer Justice
Our Subscription to Addiction
People Power United
Public Knowledge
Rinascimento Green
ReThink Citizens
SAG-AFTRA
Stand.earth
Techtonic Justice
The Oversight Lab
The Tech Oversight Project
The Value Alliance
The Leadership Conference on Civil and Human Rights
Transparency Coalition
UltraViolet
UnidosUS
Women’s Media Center
CC:
U.S. House of Representatives Oversight Committee
U.S. Senate Oversight Committee
U.S. Federal Trade Commission