Meta has confirmed that all publicly shared text and photos by adult users on Facebook and Instagram since 2007 have been utilized for training its artificial intelligence models. This acknowledgment came after initial denials during a government inquiry in Australia. The company maintains that unless users set their posts to private, their public data is subject to scraping for AI training purposes. While users in Europe have the option to opt out due to privacy laws, the same protections are not guaranteed for users in Australia or other regions, leading to concerns over data exploitation, especially involving minors.
| Signal | Change | 10y horizon | Driving force |
|---|---|---|---|
| Meta uses old public posts for AI training | From user privacy to data extraction | More stringent data privacy regulations | Monetization of user data for AI advancement |
| Users have limited control over old data | From passive sharing to data ownership | Increased user awareness and control of data | User demand for greater privacy protection |
| Regulatory differences impact data usage | From fragmented privacy rights to stricter laws | Global harmonization of data protection laws | Responses to consumer privacy concerns |
| Scraping policy affects minors retroactively | From unregulated to responsible use | Companies held accountable for past actions | Ethical considerations in data handling |
| Meta’s vague data usage policies raise concerns | From transparency to obfuscation | Better transparency in AI training practices | Public demand for clarity on data usage |