Meta has removed a controversial "AI info" setting from its platforms that previously allowed users to opt out of having their content used to train its artificial intelligence models. The company confirmed it is updating its approach to transparency and data usage for generative AI features, following pressure from privacy regulators and advocacy groups regarding how European and global user data is processed.
The Shift in Meta’s AI Data Policy
Meta’s previous policy allowed users to submit an objection form to prevent their public posts, photos, and captions from being included in the training datasets for its large language models and image generation tools. According to an official update from Meta, this "AI info" setting has been replaced by a more unified approach to managing how AI learns from public data.
The company stated that it will now rely on its legitimate interests to process public information from Facebook and Instagram to train its AI systems. This transition marks a departure from the previous opt-out mechanism that was particularly prominent in the European Union and the United Kingdom, where the company faced significant scrutiny from data protection authorities.
Regulatory Pressure and Privacy Concerns
The decision to remove the opt-out setting follows direct intervention from the Irish Data Protection Commission (DPC), which acts as Meta’s lead privacy regulator in the EU. In June 2024, the DPC requested that Meta pause its plans to train its AI models using public content shared by adults on Facebook and Instagram in the EU and EEA.
According to a press release from the Irish DPC, the regulator welcomed the move as a necessary step to ensure compliance with the General Data Protection Regulation (GDPR). Privacy advocacy group NOYB (None of Your Business) had previously filed complaints across multiple European countries, arguing that Meta’s initial plan to process user data for AI training lacked a valid legal basis under European law.
How Meta Uses Public Content
Meta maintains that it only uses public information for training purposes. This includes public posts, comments, photos, and captions. The company explicitly states that it does not use private messages between friends and family to train its AI.
For users concerned about their data footprint, Meta’s current documentation notes that even with the removal of the specific "AI info" toggle, the company continues to provide tools for users to access, rectify, or delete their information. However, the automated objection form that previously allowed users to blanket-opt out of AI training has been deprecated in favor of this new, broader policy framework.
Key Takeaways
- Policy Change: Meta has removed the specific "AI info" setting that allowed users to opt out of AI training.
- Legal Basis: The company is shifting its data processing justification to "legitimate interests" for training generative AI models.
- Regulatory Oversight: The move follows a June 2024 pause requested by the Irish Data Protection Commission regarding the use of European user data.
- Scope: The training data remains limited to public posts and photos; private messages and content from users under 18 are excluded from these specific training sets.
As Meta continues to integrate AI across its ecosystem, the tension between rapid feature deployment and international privacy standards remains a primary challenge. The company has indicated it will continue to engage with regulators to refine its transparency efforts as it develops its generative AI capabilities.