A viral “Goodbye Meta AI” copypasta has been spreading across Instagram Stories, posted by celebrities including Tom Brady, Cat Power, and James McAvoy, claiming to prevent Meta from using users’ data to train artificial intelligence models. However, Meta has confirmed these statements are not legally binding and have no effect on the company’s data usage policies.
The copypasta—filled with pseudo-legal language—represents the latest iteration of similar viral posts that have circulated on Facebook since at least 2012. Business Insider and Facebook’s PR department have debunked these types of posts multiple times over the years, including in 2015, 2016, and 2019. Despite their obvious lack of legal validity, the current wave reflects genuine user concerns about AI training practices.
Meta is indeed using public Facebook posts and photos dating back to 2007 to train its AI models, including its Llama language model and Meta AI assistant. Critically, there is no opt-out option available for users outside the European Union, where regulations require such provisions. The only alternative for users concerned about their data being used for AI training is to set all their posts to private.
In an interview with The Verge’s Alex Heath during Meta Connect, CEO Mark Zuckerberg acknowledged the concerns but stopped short of offering concrete solutions. When asked if he sympathized with creators who feel their data is being used without compensation or control, Zuckerberg suggested that concepts around fair use and data ownership “are basically going to need to get relitigated and rediscussed in the AI era.”
The viral spread of this copypasta—despite its obvious ineffectiveness—highlights a significant trust deficit between Meta and its users. The company has faced numerous privacy controversies over the years, contributing to user skepticism about how their content is being utilized. Rather than dismissing these posts as mere misinformation, the phenomenon reveals deeper anxieties about AI’s impact on personal data, creative ownership, and the power imbalance between tech platforms and their users.
Meta’s approach of using nearly two decades of user-generated content for AI training without meaningful opt-out options (except where legally required) has created friction with its user base, even as the company positions itself as a leader in open-source AI development.
Key Quotes
I think that in any new medium in technology, there are the concepts around fair use and where the boundary is between what you have control over. When you put something out in the world, to what degree do you still get to control it and own it and license it? I think that all these things are basically going to need to get relitigated and rediscussed in the AI era.
Meta CEO Mark Zuckerberg responded to questions about user concerns regarding AI training data during an interview with The Verge. His response suggests the company views data ownership as an unsettled legal question rather than acknowledging users’ immediate concerns about consent and compensation.
Meta is indeed using public Facebook posts and photos going back as far as 2007 for training its AI models. There is no opt-out, either (except in the EU, where regulations require it) other than setting your posts to private.
This statement confirms the reality behind users’ concerns—Meta is actively using nearly two decades of user content for AI training without providing opt-out mechanisms outside regulated jurisdictions, validating the anxieties driving the viral copypasta phenomenon.
People’s willingness to fall for these copypastas isn’t a sign that people are rubes; it’s a sign of how Facebook has sometimes failed to earn people’s trust over the years.
The article’s analysis highlights that the viral spread of legally meaningless text reflects deeper trust issues stemming from Meta’s history of privacy controversies, rather than user gullibility.
Our Take
This incident reveals a fundamental disconnect in how AI companies and users view data ownership. Meta’s technically correct position—that the copypasta has no legal effect—misses the larger point: users feel powerless and exploited. The EU-only opt-out provision demonstrates that Meta can provide user controls when legally required, making the absence of such options elsewhere appear like a choice rather than a technical limitation. Zuckerberg’s vague promise to “relitigate” fair use concepts in the AI era offers little comfort to users whose content is being used right now. This controversy foreshadows larger battles ahead as generative AI’s hunger for training data collides with creator rights and privacy expectations. Companies that proactively address these concerns with transparent policies and meaningful consent mechanisms will likely fare better than those waiting for regulatory mandates.
Why This Matters
This controversy illuminates a critical tension in the AI industry: the need for massive training datasets versus users’ rights to control their own content. As AI companies race to develop more sophisticated models, they’re increasingly relying on publicly available user-generated content, often without explicit consent or compensation.
Meta’s stance—that public posts are fair game for AI training—may become a precedent-setting approach that other social media platforms follow. The fact that opt-out options are only available in the EU demonstrates how regulation can protect user rights, potentially influencing future AI legislation in other jurisdictions.
The viral spread of ineffective copypastas, even among celebrities, reveals a significant communication gap between tech companies and users about AI practices. This trust deficit could impact user engagement, platform growth, and ultimately shape public opinion on AI regulation. For businesses developing AI models, this serves as a warning: transparency and user consent aren’t just ethical considerations—they’re essential for maintaining public trust and avoiding regulatory backlash.
Recommended Reading
For those interested in learning more about artificial intelligence, machine learning, and effective AI communication, here are some excellent resources:
Recommended Reading
Related Stories
- Photobucket is licensing your photos and images to train AI without your consent, and there’s no easy way to opt out
- Meta’s AI advisory council is overwhelmingly white and male, raising concerns about bias
- Meta’s Nick Clegg and Joel Kaplan to step down from key roles
- Jenna Ortega Speaks Out Against Explicit AI-Generated Images of Her