Meta Platforms has announced its decision to resume data scraping plans in the UK, focusing on utilizing public Facebook and Instagram user posts to train generative AI models [4].
Description
The company will scrape public posts, comments [1] [4] [5], photos [4], and captions to enhance its GenAI models [4], excluding private messages and content from British users under 18 [3] [4]. Users will have their data fed into Meta’s AI products without consent and will need to opt out instead [5]. Meta has integrated regulatory feedback into a revised “opt-out” approach to ensure transparency [1], with in-app notifications explaining the data collection to be rolled out to UK users starting next week. The Information Commissioner’s Office is closely monitoring the project [4], emphasizing the importance of transparency and safeguards when using personal data for model training [4]. While Meta’s plans remain on hold in the EU and EEA due to lack of regulatory approval from the ICO, the company has faced challenges in Europe over privacy regulations [1], with complaints filed against its use of an opt-out mechanism for data processing [1]. Open Rights Group has criticized the ICO for not adequately protecting the data rights of Meta users in the UK and has called for an investigation into Meta’s plans [5]. Despite doubts from privacy experts [1], Meta has relied on the legal basis of “legitimate interest” to justify its actions [1]. The decision to restart AI training efforts in the UK rather than the EU is significant [1], given the differences in regulatory enforcement between the two regions [1]. The ICO issued a statement regarding Meta’s announcement to use user data for AI training [2]. Meta paused its plans in June 2024 but has now resumed them [2]. The ICO will monitor the situation and Meta must ensure ongoing compliance [2]. Organizations using user data for AI training must be transparent and provide safeguards for users to object to the processing [2].
Conclusion
The decision to resume data scraping in the UK has raised concerns about privacy and data protection. Meta must address these concerns and ensure compliance with regulations to protect user rights. The ongoing monitoring by the ICO highlights the importance of transparency and accountability in AI training efforts. Moving forward, organizations must prioritize user consent and data protection to build trust and maintain ethical practices in AI development.
References
[1] https://techcrunch.com/2024/09/13/meta-reignites-plans-to-train-ai-using-uk-users-public-facebook-and-instagram-posts/
[2] https://www.dataguidance.com/news/uk-ico-issues-statement-metas-announcement-user-data
[3] https://www.channelnewsasia.com/business/meta-start-using-public-posts-facebook-instagram-uk-train-ai-4607341
[4] https://www.infosecurity-magazine.com/news/meta-controversial-ai-training-uk/
[5] https://www.openrightsgroup.org/press-releases/ico-fails-uk-meta-users-and-allows-social-media-giant-to-resume-data-scraping-for-ai/
