Meta, the parent company of Facebook and Instagram, faced criticism over its plans to use publicly shared user content to develop and improve its artificial intelligence (AI) models. The proposed changes raised concerns among privacy advocates, social media users and digital rights organisations about how personal information can be collected and used by technology companies.
The controversy gained attention in June 2024, when Meta notified users in the United Kingdom and Europe about changes to its privacy policy. Under the proposed approach, publicly shared posts, photographs, captions, comments and Stories from adult users could be used to improve Meta’s AI products.
Although Meta said the policy was intended to make its AI technology more relevant to different languages and cultures, critics questioned whether users had sufficient control over their information.
The issue also highlighted a wider debate in the technology industry: Who owns the content people publish online, and should companies be allowed to use it to train commercial AI systems?
What Is Meta’s New AI Data Policy?
Meta’s proposed privacy policy changes were designed to allow the company to use publicly available content shared by adult users on Facebook and Instagram to develop and improve its AI systems.
The information covered by the original proposal included:
- Public Facebook and Instagram posts.
- Photographs and videos shared publicly.
- Image captions and descriptions.
- Public comments.
- Instagram and Facebook Stories shared with a public audience.
However, Meta said it would not use private messages exchanged between users with friends and family to train its AI models. The original policy also excluded content from accounts belonging to people under 18.
The company explained that publicly available information could help its AI systems understand different languages, regional expressions, cultural references and local interests.
The aim was to make AI-generated responses and experiences more relevant to people in different regions.
However, the proposed changes raised questions about whether making content public on a social media platform automatically means users agree to its use in AI development.
Why Is Meta Using Facebook and Instagram Posts for AI Training?
Training artificial intelligence models requires large amounts of information. AI systems analyse patterns in text, images and other data to learn how to generate responses and produce content.
Meta operates several AI products, including Meta AI and its Llama family of large language models.
According to Meta, publicly shared content can help these systems understand how people communicate in different parts of the world.
For example, public posts and comments may contain regional expressions, local references, cultural discussions and language variations that are difficult to capture through generic datasets.
Meta believes using this information can improve the relevance of its AI experiences.
The company also pointed to the importance of building AI systems that reflect the languages and cultures of the communities using them.
However, critics argue that the usefulness of publicly available content does not automatically resolve questions about consent, privacy and ownership.
Why Are Privacy Organisations Criticising Meta?
Meta’s policy attracted criticism from digital rights organisations, particularly because the original approach relied on an opt-out process rather than asking users to provide explicit consent before their public content was used.
One of the main organisations challenging the proposal was NOYB (None of Your Business), a European digital rights group founded by privacy activist Max Schrems.
NOYB filed complaints with 11 data protection authorities across Europe, asking regulators to intervene and prevent Meta from proceeding with its plans.
The organisation argued that Meta’s approach raised concerns under European data protection rules, particularly the General Data Protection Regulation (GDPR).
Privacy advocates also questioned whether users could make an informed decision when the notification and objection process required them to navigate a form and explain how the proposed data processing affected them.
The criticism was not limited to whether Meta could technically access public information. It also concerned whether users had a meaningful opportunity to control how their content was used.
How Does Meta’s AI Data Opt-Out Process Work?
Under the original UK and European proposal, Meta relied on a legal basis known as legitimate interests to process certain publicly shared user information.
This meant users who did not want their information used for AI development had to exercise their right to object rather than simply waiting for a consent request.
The original objection process involved accessing a form through the notification or privacy information provided by Meta.
Users were asked to explain why they objected to the processing of their information.
Meta stated that it would review objections under applicable data protection laws and stop using information where the objection was upheld, unless it could demonstrate compelling legitimate grounds that outweighed the individual’s rights and interests.
This approach became a major point of criticism because privacy advocates believed users should not have to take additional steps to prevent their content from being used.
It is important to note that the original process and subsequent objection arrangements have changed over time. Users should check the current privacy controls and notices available in their region rather than relying on older instructions.
Can Meta Use Your Content Even If You Do Not Have an Account?
One of the more complicated aspects of the original policy concerned information about people who do not have Facebook or Instagram accounts.
Meta explained that even people without an account, or users who successfully objected, might still appear in publicly shared content uploaded by someone else.
For example, a person who does not use Instagram could appear in a photograph posted publicly by another user.
In such cases, information about that person may still be present in the publicly shared content.
This raises additional privacy questions because individuals may have limited control over photographs, videos or other information shared by other people.
It also demonstrates that managing personal data online is not always as simple as adjusting the privacy settings of one’s own social media account.
What Did Meta Say in Response to the Criticism?
Meta defended its approach by arguing that using publicly shared information was important for developing AI products that understand different communities.
In its June 2024 explanation, the company said it was following approaches used by other major technology companies and believed its proposed process complied with applicable European privacy laws.
Meta also highlighted several safeguards:
- Private messages between friends and family would not be used for AI training.
- Content from accounts belonging to users under 18 would be excluded from the original European proposal.
- Users would be notified about the planned processing.
- People would have a way to object to the use of their information.
Meta argued that its AI models learn patterns from information rather than functioning as databases designed to identify individual users.
However, these explanations did not remove concerns among privacy campaigners, who continued to question the legal basis and effectiveness of the objection process.
What Happened After the Original Announcement?
Meta’s plans did not proceed exactly as originally announced.
In June 2024, the company paused its planned use of public Facebook and Instagram content for AI training in the European region following intervention from the Irish Data Protection Commission, which leads important aspects of Meta’s EU data protection oversight.
The UK position later changed. In September 2024, Meta announced that it would move ahead with using publicly shared content from adult Facebook and Instagram users in the UK after discussions with the Information Commissioner’s Office (ICO).
The ICO clarified that it had not given Meta regulatory approval. It said the company remained responsible for demonstrating compliance with data protection law and providing effective safeguards.
In April 2025, Meta also announced that it would resume AI training using public content from adult users in the European Union. The company said it would notify users and provide an opportunity to object.
According to subsequent reporting, the resumed approach included public posts and comments, as well as certain interactions with Meta AI.
These developments are important because the June 2024 BBC report describes the original proposal and initial regulatory dispute, not the complete current position.
What Does This Mean for User Privacy?
Meta’s policy controversy highlights several privacy issues that extend beyond Facebook and Instagram.
1. Public content does not remove every privacy concern
Users may share information publicly without expecting it to become part of a commercial AI training dataset. The distinction between public visibility and permission for additional uses remains important.
2. Users need clear information
Privacy policies should explain what information is collected, why it is processed and how people can exercise their rights.
3. Opt-out systems must be accessible
If a company relies on an objection mechanism, users should be able to find and understand it without unnecessary complexity.
4. Personal information can appear in other people’s content
A person may be included in a public photograph, video or post even if they do not maintain an account on the platform.
5. AI development raises new questions about content rights
The use of online content to train AI models creates ongoing debates about copyright, consent, ownership and fair compensation for creators.
How Can Facebook and Instagram Users Protect Their Privacy?
Users can take several practical steps to manage the information they share on social media.
- Review Facebook and Instagram privacy settings regularly.
- Avoid sharing sensitive personal information publicly.
- Check Meta’s current Privacy Policy and AI-related notices.
- Review available objection options in the relevant account settings.
- Be careful when sharing photographs or personal information about other people.
- Review the audience settings of older posts and public content.
Users in different countries may have different privacy rights and objection options depending on local regulations.
Changing an account to private can help restrict future audience access, but it does not automatically remove information that has already been shared publicly or guarantee that previously collected data will be excluded from every form of processing.
Conclusion
Meta’s plans to use public Facebook and Instagram content for AI development created an important debate about privacy, consent and digital ownership.
The company argues that publicly shared information helps improve AI systems and makes them more relevant to different communities. However, privacy advocates believe users should have meaningful control over how their content is collected and used.
The controversy also demonstrates that social media privacy policies are becoming increasingly important as technology companies expand their AI capabilities.
As artificial intelligence continues to develop, companies will face growing pressure to explain their data practices, respect applicable privacy laws and provide transparent choices to users.
For social media users, the key lesson is simple: sharing something publicly does not mean every possible future use of that information is automatically clear or understood.
