Meta has confirmed that all text and photos publicly shared by adult users on Facebook and Instagram since 2007 have been utilized to train its artificial intelligence models. During a local government inquiry on AI adoption, Melinda Claybaugh, Meta’s global privacy director, initially denied claims about the use of user data from 2007 but later acknowledged it under further questioning.
Green Party senator David Shoebridge highlighted the situation, stating, “Unless you have consciously set those posts to private since 2007, Meta has decided to scrape all public posts on Instagram and Facebook.” Claybaugh conceded this was accurate.
Meta's privacy center and blog state that they use public posts and comments from Facebook and Instagram to train generative AI models for various features and the open-source community. However, the company has been vague regarding how data is collected, when it began scraping, and the extent of this collection. In June, when questioned by The New York Times, Meta confirmed that setting posts to "private" would stop future scraping, but it would not erase already collected data. This poses issues for individuals who shared posts back in 2007, particularly minors unaware their data could be used in this manner.
Claybaugh asserted that Meta does not scrape data from users under 18. However, when asked if public photos of his children would be scraped from Senator Tony Sheldon’s account, she confirmed that they would be, but she could not clarify if accounts created as minors were also subject to scraping.
While European users have the option to opt-out due to stringent privacy regulations, billions of users in other regions, including Australia, currently lack this choice if they want to keep their posts public. Claybaugh indicated that Australian users might not receive future opt-out options, explaining that European regulations prompted their ability to protect user data. Shoebridge responded by criticizing the Australian government's inaction on privacy, stating, "The failure to act means companies like Meta continue to exploit pictures and videos of children on Facebook."